The simplest LogsQL query is just a [word](#word), which must be found in the [log message](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field).
The `AND` operator means that the [log entry](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) must match both filters in order to be selected.
Typical LogsQL query consists of multiple [filters](#filters) joined with `AND` operator. It may be tiresome typing and then reading all these `AND` words.
The query returns logs in arbitrary order because sorting of big amounts of logs may require non-trivial amounts of CPU and RAM.
The number of logs with `error` word over the last 5 minutes isn't usually too big (e.g. less than a few millions), so it is OK to sort them with [`sort` pipe](#sort-pipe).
The following query sorts the selected logs by [`_time`](https://docs.victoriametrics.com/victorialogs/keyconcepts/#time-field) field:
```logsql
_time:5m error | sort by (_time)
```
It is unlikely you are going to investigate more than a few hundreds of logs returned by the query above. So you can limit the number of returned logs
with [`limit` pipe](#limit-pipe). The following query returns the last 10 logs with the `error` word over the last 5 minutes:
```logsql
_time:5m error | sort by (_time) desc | limit 10
```
By default VictoriaLogs returns all the [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
If you need only the given set of fields, then add [`fields` pipe](#fields-pipe) to the end of the query. For example, the following query returns only
Suppose the query above selects too many rows because some buggy app pushes invalid error logs to VictoriaLogs. Suppose the app adds `buggy_app` [word](#word) to every log line.
Then the following query removes all the logs from the buggy app, allowing us paying attention to the real errors:
This query uses `NOT` [operator](#logical-filter) for removing log lines from the buggy app. The `NOT` operator is used frequently, so it can be substituted with `!` char
(the `!` char is used instead of `-` char as a shorthand for `NOT` operator because it nicely combines with [`=`](https://docs.victoriametrics.com/victorialogs/logsql/#exact-filter)
The parentheses are **required** here, since otherwise the query won't return the expected results.
The query `error !buggy_app OR foobar` is interpreted as `(error AND NOT buggy_app) OR foobar` according to [priorities for AND, OR and NOT operator](#logical-filters).
This query returns logs with `foobar` [word](#word), even if do not contain `error` word or contain `buggy_app` word.
So it is recommended wrapping the needed query parts into explicit parentheses if you are unsure in priority rules.
Queries above assume that the `error` [word](#word) is stored in the [log message](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field).
If this word is stored in other [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) such as `log.level`, then add `log.level:` prefix
What if the application identifier - such as `buggy_app` and `foobar` - is stored in the `app` field? Correct - just add `app:` prefix in front of `buggy_app` and `foobar`:
In this case it is recommended associating the `app` field with [log stream fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#stream-fields)
during [data ingestion](https://docs.victoriametrics.com/victorialogs/data-ingestion/). This usually improves both compression rate
This query skips scanning for [log messages](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field) from `buggy_app` and `foobar` apps.
It inpsects only `log.level` and [`_stream`](https://docs.victoriametrics.com/victorialogs/keyconcepts/#stream-fields) labels.
This significantly reduces disk read IO and CPU time needed for performing the query.
LogsQL also provides [functions for statistics calculation](#stats-pipe) over the selected logs. For example, the following query returns the number of logs
Now you are familiar with LogsQL basics. See [LogsQL examples](https://docs.victoriametrics.com/victorialogs/logsql-examples/) and [query syntax](#query-syntax)
Additionally to filters, LogQL query may contain arbitrary mix of optional actions for processing the selected logs. These actions are delimited by `|` and are known as [`pipes`](#pipes).
For example, the following query uses [`stats` pipe](#stats-pipe) for returning the number of [log messages](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field)
with the `error` [word](#word) for the last 5 minutes:
then its' name followed by the colon must be put in front of the filter. For example, if `error` [word filter](#word-filter) must be applied
to the `log.level` field, then use `log.level:error` query.
Field names and filter args can be put into quotes if they contain special chars, which may clash with LogsQL syntax. LogsQL supports quoting via double quotes `"`,
single quotes `'` and backticks:
```logsql
"some 'field':123":i('some("value")') AND `other"value'`
```
If doubt, it is recommended quoting field names and filter args.
- [Time filter](#time-filter) - matches logs with [`_time` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#time-field) in the given time range
- [Day range filter](#day-range-filter) - matches logs with [`_time` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#time-field) in the given per-day time range
- [Week range filter](#week-range-filter) - matches logs with [`_time` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#time-field) in the given per-week day range
- [Stream filter](#stream-filter) - matches logs, which belong to the given [streams](https://docs.victoriametrics.com/victorialogs/keyconcepts/#stream-fields)
- [Empty value filter](#empty-value-filter) - matches logs without the given [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
- [Any value filter](#any-value-filter) - matches logs with the given non-empty [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
- [Range filter](#range-filter) - matches logs with numeric [field values](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) in the given range
- [IPv4 range filter](#ipv4-range-filter) - matches logs with ip address [field values](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) in the given range
- [String range filter](#string-range-filter) - matches logs with [field values](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) in the given string range
- [Length range filter](#length-range-filter) - matches logs with [field values](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) of the given length range
VictoriaLogs scans all the logs per each query if it doesn't contain the filter on [`_time` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#time-field).
-`_time:duration` matches logs with timestamps on the time range `(now-duration, now]`, where `duration` can have [these values](#duration-values). Examples:
-`_time:YYYY-MM-DD` - matches all the logs for the particular day by UTC. For example, `_time:2023-04-25` matches logs on April 25, 2023 by UTC.
-`_time:YYYY-MM` - matches all the logs for the particular month by UTC. For example, `_time:2023-02` matches logs on February, 2023 by UTC.
-`_time:YYYY` - matches all the logs for the particular year by UTC. For example, `_time:2023` matches logs on 2023 by UTC.
-`_time:YYYY-MM-DDTHH` - matches all the logs for the particular hour by UTC. For example, `_time:2023-04-25T22` matches logs on April 25, 2023 at 22 hour by UTC.
-`_time:YYYY-MM-DDTHH:MM` - matches all the logs for the particular minute by UTC. For example, `_time:2023-04-25T22:45` matches logs on April 25, 2023 at 22:45 by UTC.
-`_time:YYYY-MM-DDTHH:MM:SS` - matches all the logs for the particular second by UTC. For example, `_time:2023-04-25T22:45:59` matches logs on April 25, 2023 at 22:45:59 by UTC.
- It is recommended specifying the smallest possible time range during the search, since it reduces the amounts of log entries, which need to be scanned during the query.
`_time:day_range[start, end]` filter allows returning logs on the particular `start ... end` time per every day, where `start` and `end` have the format `hh:mm`.
For example, the following query matches logs between `08:00` and `18:00` UTC every day:
```logsql
_time:day_range[08:00, 18:00)
```
This query includes `08:00`, while `18:00` is excluded, e.g. the last matching time is `17:59:59.999999999`.
Replace `[` with `(` in order to exclude the starting time. Replace `)` with `]` in order to include the ending time.
For example, the following query matches logs between `08:00` and `18:00`, excluding `08:00:00.000000000` and including `18:00`:
```logsql
_time:day_range(08:00, 18:00]
```
If the time range must be applied to other than UTC time zone, then add `offset <duration>`, where `<duration>` can have [any supported duration value](#duration-values).
For example, the following query selects logs between `08:00` and `18:00` at `+0200` time zone:
```logsql
_time:day_range[08:00, 18:00) offset 2h
```
Performance tip: it is recommended specifying regular [time filter](#time-filter) additionally to `day_range` filter. For example, the following query selects logs
between `08:00` and `20:00` every day for the last week:
```logsql
_time:1w _time:day_range[08:00, 18:00)
```
See also:
- [Week range filter](#week-range-filter)
- [Time filter](#time-filter)
### Week range filter
`_time:week_range[start, end]` filter allows returning logs on the particular `start ... end` days per every day, where `start` and `end` can have the following values:
-`Sun` or `Sunday`
-`Mon` or `Monday`
-`Tue` or `Tuesday`
-`Wed` or `Wednesday`
-`Thu` or `Thusday`
-`Fri` or `Friday`
-`Sat` or `Saturday`
For example, the following query matches logs between Monday and Friday UTC every day:
```logsql
_time:week_range[Mon, Fri]
```
This query includes Monday and Friday.
Replace `[` with `(` in order to exclude the starting day. Replace `]` with `)` in order to exclude the ending day.
For example, the following query matches logs between Sunday and Saturday, excluding Sunday and Saturday (e.g. it is equivalent to the previous query):
```logsql
_time:week_range(Sun, Sat)
```
If the day range must be applied to other than UTC time zone, then add `offset <duration>`, where `<duration>` can have [any supported duration value](#duration-values).
For example, the following query selects logs between Monday and Friday at `+0200` time zone:
```logsql
_time:week_range[Mon, Fri] offset 2h
```
The `week_range` filter can be combined with [`day_range` filter](#day-range-filter) using [logical filters](#logical-filter). For example, the following query
selects logs between `08:00` and `18:00` every day of the week excluding Sunday and Saturday:
Performance tip: it is recommended specifying regular [time filter](#time-filter) additionally to `week_range` filter. For example, the following query selects logs
between Monday and Friday per every week for the last 4 weeks:
VictoriaLogs provides an optimized way to select logs, which belong to particular [log streams](https://docs.victoriametrics.com/victorialogs/keyconcepts/#stream-fields).
Every [log stream](https://docs.victoriametrics.com/victorialogs/keyconcepts/#stream-fields) in VictoriaMetrics is uniquely identified by `_stream_id` field.
The `_stream_id:...` filter allows quickly selecting all the logs belonging to the particular stream.
For example, the following query selects all the logs, which belong to the [log stream](https://docs.victoriametrics.com/victorialogs/keyconcepts/#stream-fields)
with `_stream_id` equal to `0000007b000001c850d9950ea6196b1a4812081265faa1c7`:
If the log stream contains too many logs, then it is good idea limiting the number of returned logs with [time filter](#time-filter). For example, the following
query selects logs for the given stream for the last hour:
-`multiple errors occurred`, since the `errors` word doesn't match `error` word. Use `error*` for this case. See [these docs](#prefix-filter) for details.
if it must be searched in the given field. For example, the following query returns log entries containing the `error` [word](#word) in the `log.level` field:
```logsql
log.level:error
```
Both the field name and the word in the query can contain arbitrary [utf-8](https://en.wikipedia.org/wiki/UTF-8)-encoded chars. For example:
```logsql
поле:значение
```
Both the field name and the word in the query can be put inside quotes if they contain special chars, which may clash with the query syntax.
For example, the following query searches for the ip `1.2.3.45` in the field `ip:remote`:
```logsql
"ip:remote":"1.2.3.45"
```
See also:
- [Phrase filter](#phrase-filter)
- [Exact filter](#exact-filter)
- [Prefix filter](#prefix-filter)
- [Logical filter](#logical-filter)
### Phrase filter
Is you need to search for log messages with the specific phrase inside them, then just wrap the phrase in quotes.
The phrase can contain any chars, including whitespace, punctuation, parens, etc. They are taken into account during the search.
-`ssh login fail`, since the message misses `:` char just after the `ssh`.
Use `seq("ssh", "login", "fail")` query if log messages with the sequence of these words must be found. See [these docs](#sequence-filter) for details.
-`login fail: ssh error`, since the message doesn't contain the full phrase requested in the query. If you need matching a message
with all the [words](#word) listed in the query, then use `ssh AND login AND fail` query. See [these docs](#logical-filter) for details.
-`ssh: login failed`, since the message ends with `failed` [word](#word) instead of `fail` word. Use `"ssh: login fail"*` query for this case.
If the phrase contains double quotes, then either put `\` in front of double quotes or put the phrase inside single quotes. For example, the following filter searches
if it must be searched in the given field. For example, the following query returns log entries containing the `cannot open file` phrase in the `event.original` field:
```logsql
event.original:"cannot open file"
```
Both the field name and the phrase can contain arbitrary [utf-8](https://en.wikipedia.org/wiki/UTF-8)-encoded chars. For example:
```logsql
сообщение:"невозможно открыть файл"
```
The field name can be put inside quotes if it contains special chars, which may clash with the query syntax.
For example, the following query searches for the `cannot open file` phrase in the field `some:message`:
```logsql
"some:message":"cannot open file"
```
See also:
- [Exact filter](#exact-filter)
- [Word filter](#word-filter)
- [Prefix filter](#prefix-filter)
- [Logical filter](#logical-filter)
### Prefix filter
If you need to search for log messages with [words](#word) / phrases containing some prefix, then just add `*` char to the end of the [word](#word) / phrase in the query.
For example, the following query returns [log messages](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field), which contain [words](#word) with `err` prefix:
This query doesn't match the following log messages:
-`Error: foobar`, since the `Error` [word](#word) starts with capital letter. Use `i(err*)` for this case. See [these docs](#case-insensitive-filter) for details.
If the prefix contains double quotes, then either put `\` in front of double quotes or put the prefix inside single quotes. For example, the following filter searches
in order to apply it to the given field. For example, the following query matches `log.level` field containing any word with the `err` prefix:
```logsql
log.level:err*
```
If the field name contains special chars, which may clash with the query syntax, then it may be put into quotes in the query.
For example, the following query matches `log:level` field containing any word with the `err` prefix.
```logsql
"log:level":err*
```
Performance tips:
- Prefer using [word filters](#word-filter) and [phrase filters](#phrase-filter) combined via [logical filter](#logical-filter)
instead of prefix filter.
- Prefer moving [word filters](#word-filter) and [phrase filters](#phrase-filter) in front of prefix filter when using [logical filter](#logical-filter).
- See [other performance tips](#performance-tips).
LogsQL supports `field:>X`, `field:>=X`, `field:<X` and `field:<=X` filters, where `field` is the name of [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
and `X` is [numeric value](#numeric-values), IPv4 address or a string. For example, the following query returns logs containing numeric values for the `response_size` field bigger than `10*1024`:
Sometimes it is needed to find log entries containing any non-empty value for the given [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
The [word filter](#word-filter) and [phrase filter](#phrase-filter) return [log messages](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field),
which contain the given word or phrase inside them. The message may contain additional text other than the requested word or phrase. If you need searching for log messages
other than the specified in the `exact` filter. Use `"fatal error: cannot find /foo/bar"` query in this case. See [these docs](#phrase-filter) for details.
By default the `exact` filter is applied to the [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field).
Specify the [field name](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) in front of the `exact` filter and put a colon after it
By default the `exact` filter is applied to the [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field).
Specify the [field name](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) in front of the `exact` filter and put a colon after it
if it must be searched in the given field. For example, the following query returns log entries with `log.level` field, which starts with `err` prefix:
Sometimes it is needed to locate log messages with a field containing one of the given values. This can be done with multiple [exact filters](#exact-filter)
-`FooError`, since the `FooError` [word](#word) has superfluous prefix `Foo`. Use `~"(?i)error"` for this case. See [these docs](#regexp-filter) for details.
-`too many Errors`, since the `Errors` [word](#word) has superfluous suffix `s`. Use `i(error*)` for this case.
with [words](#word) or phrases in a particular order. For example, if log messages with `error` word followed by `open file` phrase
must be found, then the following LogsQL query can be used:
```logsql
seq("error", "open file")
```
This query matches `some error: cannot open file /foo/bar` message, since the `open file` phrase goes after the `error` [word](#word).
The query doesn't match the `cannot open file: error` message, since the `open file` phrase is located in front of the `error` [word](#word).
If you need matching log messages with both `error` word and `open file` phrase, then use `error AND "open file"` query. See [these docs](#logical-filter)
The query matches the following [log messages](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field), which contain either `err` or `warn` substrings:
If the regexp contains double quotes, then either put `\` in front of double quotes or put the regexp inside single quotes. For example, the following regexp searches
- Prefer using `="some prefix"*` instead of `~"^some prefix"`, since the [`exact` filter](#exact-prefix-filter) works much faster than the regexp filter.
If you need to filter log message by some field containing only [IPv4](https://en.wikipedia.org/wiki/Internet_Protocol_version_4) addresses such as `1.2.3.4`,
then the `ipv4_range()` filter can be used. For example, the following query matches log entries with `user.ip` address in the range `[127.0.0.0 - 127.255.255.255]`:
```logsql
user.ip:ipv4_range(127.0.0.0, 127.255.255.255)
```
The `ipv4_range()` accepts also IPv4 subnetworks in [CIDR notation](https://en.wikipedia.org/wiki/Classless_Inter-Domain_Routing#CIDR_notation).
For example, the following query is equivalent to the query above:
```logsql
user.ip:ipv4_range("127.0.0.0/8")
```
If you need matching a single IPv4 address, then just put it inside `ipv4_range()`. For example, the following query matches `1.2.3.4` IP
since the `127.0.0.1` ip is surrounded by other text. Extract the IP from the message with `parse(_msg, "request from <ip>: done")` [transformation](#transformations)
- If you need searching for [log messages](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field) containing the given `X.Y.Z.Q` IPv4 address,
at least a single IPv4 address out of the given list, then `"ip1" OR "ip2" ... OR "ipN"` query can be used. See [these docs](#logical-filter) for details.
- If you need finding log entries with `ip` field in multiple ranges, then use `ip:(ipv4_range(range1) OR ipv4_range(range2) ... OR ipv4_range(rangeN))` query.
It is possible to use `inf` as the upper bound. For example, the following query matches [log messages](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field)
the filter to the needed field. For example, the following query matches log entries with the `foo` field length in the range `[10, 20]` chars:
```logsql
foo:len_range(10, 20)
```
See also:
- [Range filter](#range-filter)
- [Logical filter](#logical-filter)
### Logical filter
Simpler LogsQL [filters](#filters) can be combined into more complex filters with the following logical operations:
-`q1 AND q2` - matches common log entries returned by both `q1` and `q2`. Arbitrary number of [filters](#filters) can be combined with `AND` operation.
which do not contain `info` [word](#word). The `NOT` operation is frequently used in LogsQL queries, so it is allowed substituting `NOT` with `!` in queries.
For example, `!info` is equivalent to `NOT info`.
The `NOT` operation has the highest priority, `AND` has the middle priority and `OR` has the lowest priority.
The priority order can be changed with parentheses. For example, `NOT info OR debug` is interpreted as `(NOT info) OR debug`,
By default logical filters apply to the [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field)
unless the inner filters explicitly specify the needed [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) via `field_name:filter` syntax.
For example, `log.level:error OR log.level:warning OR log.level:info` can be substituted with the shorter query: `log.level:(error OR warning OR info)`.
Performance tips:
- VictoriaLogs executes logical operations from the left to the right, so it is recommended moving the most specific
and the fastest filters (such as [word filter](#word-filter) and [phrase filter](#phrase-filter)) to the left,
while moving less specific and the slowest filters (such as [regexp filter](#regexp-filter) and [case-insensitive filter](#case-insensitive-filter))
- [`extract`](#extract-pipe) extracts the specified text into the given log fields.
- [`extract_regexp`](#extract_regexp-pipe) extracts the specified text into the given log fields via [RE2 regular expressions](https://github.com/google/re2/wiki/Syntax).
- [`field_values`](#field_values-pipe) returns all the values for the given [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
If some [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) must be copied, then `| copy src1 as dst1, ..., srcN as dstN` [pipe](#pipes) can be used.
For example, the following query copies `host` field to `server` for logs over the last 5 minutes, so the output contains both `host` and `server` fields:
Multiple fields can be copied with a single `| copy ...` pipe. For example, the following query copies
[`_time` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#time-field) to `timestamp`, while [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field)
is copied to `message`:
```logsql
_time:5m | copy _time as timestmap, _msg as message
```
The `as` keyword is optional.
`cp` keyword can be used instead of `copy` for convenience. For example, `_time:5m | cp foo bar` is equivalent to `_time:5m | copy foo as bar`.
If some [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) must be deleted, then `| delete field1, ..., fieldN` [pipe](#pipes) can be used.
`drop`, `del` and `rm` keywords can be used instead of `delete` for convenience. For example, `_time:5m | drop host` is equivalent to `_time:5m | delete host`.
`| drop_empty_fields` pipe drops [fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) with empty values. It also skips log entries with zero non-empty fields.
For example, the following query drops possible empty `email` field generated by [`extract` pipe](#extract-pipe) if the `foo` field doesn't contain email:
```logsql
_time:5m | extract 'email: <email>,' from foo | drop_empty_fields
`| extract "pattern" from field_name` [pipe](#pipes) allows extracting arbitrary text into output fields according to the [`pattern`](#format-for-extract-pipe-pattern) from the given
`| extract ...` can be useful for extracting additional fields needed for further data processing with other pipes such as [`stats` pipe](#stats-pipe) or [`sort` pipe](#sort-pipe).
For example, the following query selects logs with the `error` [word](#word) for the last day,
extracts ip address from [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field) into `ip` field and then calculates top 10 ip addresses
If the `| extract ...` pipe is applied to [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field), then the `from _msg` part can be omitted.
For example, the following query is equivalent to the previous one:
```logsql
_time:1d error | extract "ip=<ip> " | stats by (ip) count() logs | sort by (logs) desc limit 10
Add `keep_original_fields` to the end of `extract ...` when the original non-empty values of the fields mentioned in the pattern must be preserved
instead of overwriting it with the extracted values. For example, the following query extracts `<ip>` only if the original value for `ip` field is missing or is empty:
By default `extract` writes empty matching fields to the output, which may overwrite existing values. Add `skip_empty_results` to the end of `extract ...`
in order to prevent from overwriting the existing values for the corresponding fields with empty values.
For example, the following query preserves the original `ip` field value if `foo` field doesn't contain the matching ip:
```logsql
_time:5m | extract 'ip=<ip> ' from foo skip_empty_results
Performance tip: it is recommended using more specific [log filters](#filters) in order to reduce the number of log entries, which are passed to `extract`.
See [general performance tips](#performance-tips) for details.
then the matching starts from the beginning of the input text. Matching is performed sequentially according to the `pattern`. If some `textX` isn't found
in the remaining input text, then the remaining named placeholders receive empty string values and the matching finishes prematurely.
Note that the user-agent part of the log message is in double quotes. This means that it may contain special chars, including escaped double quote, e.g. `\"`.
This may break proper matching of the string in double quotes.
VictoriaLogs automatically detects quoted strings and automatically unquotes them if the first matching char in the placeholder is double quote or backtick.
So it is better to use the following `pattern` for proper matching of quoted `user_agent` string:
This is useful for extracting JSON strings. For example, the following `pattern` properly extracts the `message` JSON string into `msg` field, even if it contains special chars:
The automatic string unquoting can be disabled if needed by adding `plain:` prefix in front of the field name. For example, if some JSON array of string values must be captured
into `json_array` field, then the following `pattern` can be used:
If some special chars such as `<` must be matched by the `pattern`, then they can be [html-escaped](https://en.wikipedia.org/wiki/List_of_XML_and_HTML_character_entity_references).
An alternative approach is to add `keep_original_fields` to the end of `extract`, in order to keep the original non-empty values for the extracted fields.
For example, the following query is equivalent to the previous one:
`| extract_regexp "pattern" from field_name` [pipe](#pipes) extracts substrings from the [`field_name` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
according to the provided `pattern`, and stores them into field names according to the named fields inside the `pattern`.
The `pattern` must contain [RE2 regular expression](https://github.com/google/re2/wiki/Syntax) with named fields (aka capturing groups) in the form `(?P<capture_field_name>...)`.
Matching substrings are stored to the given `capture_field_name` [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
For example, the following query extracts ipv4 addresses from [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field)
and puts them into `ip` field for logs over the last 5 minutes:
```logsql
_time:5m | extract_regexp "(?P<ip>([0-9]+[.]){3}[0-9]+)" from _msg
```
The `from _msg` part can be omitted if the data extraction is performed from the [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field).
So the following query is equivalent to the previous one:
Add `keep_original_fields` to the end of `extract_regexp ...` when the original non-empty values of the fields mentioned in the pattern must be preserved
instead of overwriting it with the extracted values. For example, the following query extracts `<ip>` only if the original value for `ip` field is missing or is empty:
By default `extract_regexp` writes empty matching fields to the output, which may overwrite existing values. Add `skip_empty_results` to the end of `extract_regexp ...`
in order to prevent from overwriting the existing values for the corresponding fields with empty values.
For example, the following query preserves the original `ip` field value if `foo` field doesn't contain the matching ip:
```logsql
_time:5m | extract_regexp 'ip=(?P<ip>([0-9]+[.]){3}[0-9]+)' from foo skip_empty_results
If some log entries must be skipped from [`extract_regexp` pipe](#extract-pipe), then add `if (<filters>)` filter after the `extract` word.
The `<filters>` can contain arbitrary [filters](#filters). For example, the following query extracts `ip`
from [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) only
if the input [log entry](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) doesn't contain `ip` field or this field is empty:
```logsql
_time:5m | extract_regexp if (ip:"") "ip=(?P<ip>([0-9]+[.]){3}[0-9]+)"
```
An alternative approach is to add `keep_original_fields` to the end of `extract_regexp`, in order to keep the original non-empty values for the extracted fields.
For example, the following query is equivalent to the previous one:
`| field_values field_name` [pipe](#pipe) returns all the values for the given [`field_name` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
It is possible limiting the number of returned values by adding `limit N` to the end of the `field_values ...`. For example, the following query returns
up to 10 values for the field `user_id` over logs for the last 5 minutes:
If the limit is reached, then the set of returned values is random. Also the number of matching logs per each returned value is zeroed for performance reasons.
It is possible to select the given set of log fields with `| fields field1, ..., fieldN` [pipe](#pipes). For example, the following query selects only `host`
and [`_msg`](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field) fields from logs for the last 5 minutes:
For example, the following query stores `request from <ip>:<port>` text into [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field),
_time:5m | format "request from <ip>:<port>" as _msg
```
If the result of the `format` pattern is stored into [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field),
then `as _msg` part can be omitted. The following query is equivalent to the previous one:
```logsql
_time:5m | format "request from <ip>:<port>"
```
If some field values must be put into double quotes before formatting, then add `q:` in front of the corresponding field name.
For example, the following command generates properly encoded JSON object from `_msg` and `stacktrace` [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
and stores it into `my_json` output field:
```logsql
_time:5m | format '{"_msg":<q:_msg>,"stacktrace":<q:stacktrace>}' as my_json
Add `keep_original_fields` to the end of `format ... as result_field` when the original non-empty value of the `result_field` must be preserved
instead of overwriting it with the `format` results. For example, the following query adds formatted result to `foo` field only if it was missing or empty:
```logsql
_time:5m | format 'some_text' as foo keep_original_fields
Add `skip_empty_results` to the end of `format ...` if empty results shouldn't be written to the output. For example, the following query adds formatted result to `foo` field
Performance tip: it is recommended using more specific [log filters](#filters) in order to reduce the number of log entries, which are passed to `format`.
See [general performance tips](#performance-tips) for details.
If only a subset of selected logs must be processed, then `| limit N` [pipe](#pipes) can be used, where `N` can contain any [supported integer numeric value](#numeric-values).
For example, the following query returns up to 100 logs over the last 5 minutes:
```logsql
_time:5m | limit 100
```
`head` keyword can be used instead of `limit` for convenience. For example, `_time:5m | head 100` is equivalent to `_time:5m | limit 100`.
By default rows are selected in arbitrary order because of performance reasons, so the query above can return different sets of logs every time it is executed.
[`sort` pipe](#sort-pipe) can be used for making sure the logs are in the same order before applying `limit ...` to them.
`| math ...` [pipe](#pipes) performs mathematical calculations over [numeric values](#numeric-values) stored in [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
Where `exprX` is one of the supported math expressions mentioned below, while `resultNameX` is the name of the field to store the calculated result to.
The `as` keyword is optional. The result name can be omitted. In this case the result is stored to a field with the name equal to string represenation
For example, the following query divides `duration_msecs` field value by 1000, then rounds it to integer and stores the result in the `duration_secs` field:
-`max(arg1, ..., argN)` - returns the maximum value among the given `arg1`, ..., `argN`
-`min(arg1, ..., argN)` - returns the minimum value among the given `arg1`, ..., `argN`
-`round(arg)` - returns rounded to integer value for the given `arg`. The `round()` accepts optional `nearest` arg, which allows rounding the number to the given `nearest` multiple.
For example, `round(temperature, 0.1)` rounds `temperature` field to one decimal digit after the point.
Every `argX` argument in every mathematical operation can contain one of the following values:
- The name of [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model). For example, `errors_total / requests_total`.
The log field is parsed into numeric value if it contains [supported numeric value](#numeric-values). The log field is parsed into [Unix timestamp](https://en.wikipedia.org/wiki/Unix_time)
in nanoseconds if it contains [rfc3339 time](https://www.rfc-editor.org/rfc/rfc3339). The log field is parsed into `uint32` number if it contains IPv4 address.
The log field is parsed into `NaN` in other cases.
- Any [supported numeric value](#numeric-values), [rfc3339 time](https://www.rfc-editor.org/rfc/rfc3339) or IPv4 address. For example, `1MiB`, `"2024-05-15T10:20:30.934324Z"` or `"12.34.56.78"`.
- Another mathematical expression, which can be put inside `(...)`. For example, `(a + b) * c`.
The parsed time, duration and IPv4 address can be converted back to string representation after math transformations with the help of [`format` pipe](#format-pipe). For example,
the following query rounds the `request_duration` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) to seconds before converting it back to string representation:
```logsql
_time:5m | math round(request_duration, 1e9) as request_duration_nsecs | format '<duration:request_duration_nsecs>' as request_duration
If some selected logs must be skipped after [`sort`](#sort-pipe), then `| offset N` [pipe](#pipes) can be used, where `N` can contain any [supported integer numeric value](#numeric-values).
For example, the following query skips the first 100 logs over the last 5 minutes after sorting them by [`_time`](https://docs.victoriametrics.com/victorialogs/keyconcepts/#time-field):
`| pack_json as field_name` [pipe](#pipe) packs all [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) into JSON object
For example, the following query packs all the fields into JSON object and stores it into [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field)
for logs over the last 5 minutes:
```logsql
_time:5m | pack_json as _msg
```
The `as _msg` part can be omitted if packed JSON object is stored into [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field).
The following query is equivalent to the previous one:
If only a subset of labels must be packed into JSON, then it must be listed inside `fields (...)` after `pack_json`. For example, the following query builds JSON with `foo` and `bar` fields
The `pack_json` doesn't modify or delete other labels. If you do not need them, then add [`| fields ...`](#fields-pipe) after the `pack_json` pipe. For example, the following query
`| pack_logfmt as field_name` [pipe](#pipe) packs all [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) into [logfmt](https://brandur.org/logfmt) message
For example, the following query packs all the fields into [logfmt](https://brandur.org/logfmt) message and stores it
into [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field) for logs over the last 5 minutes:
```logsql
_time:5m | pack_logfmt as _msg
```
The `as _msg` part can be omitted if packed message is stored into [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field).
The following query is equivalent to the previous one:
```logsql
_time:5m | pack_logfmt
```
If only a subset of labels must be packed into [logfmt](https://brandur.org/logfmt), then it must be listed inside `fields (...)` after `pack_logfmt`.
For example, the following query builds [logfmt](https://brandur.org/logfmt) message with `foo` and `bar` fields only and stores the result in `baz` field:
```logsql
_time:5m | pack_logfmt fields (foo, bar) as baz
```
The `pack_logfmt` doesn't modify or delete other labels. If you do not need them, then add [`| fields ...`](#fields-pipe) after the `pack_logfmt` pipe. For example, the following query
leaves only the `foo` label with the original log fields packed into [logfmt](https://brandur.org/logfmt):
If some [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) must be renamed, then `| rename src1 as dst1, ..., srcN as dstN` [pipe](#pipes) can be used.
For example, the following query renames `host` field to `server` for logs over the last 5 minutes, so the output contains `server` field instead of `host` field:
```logsql
_time:5m | rename host as server
```
Multiple fields can be renamed with a single `| rename ...` pipe. For example, the following query renames `host` to `instance` and `app` to `job`:
```logsql
_time:5m | rename host as instance, app as job
```
The `as` keyword is optional.
`mv` keyword can be used instead of `rename` keyword for convenience. For example, `_time:5m | mv foo bar` is equivalent to `_time:5m | rename foo as bar`.
in the given [`field`](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
For example, the following query replaces all the `secret-password` substrings with `***` in the [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field)
for logs over the last 5 minutes:
```logsql
_time:5m | replace ("secret-password", "***") at _msg
```
The `at _msg` part can be omitted if the replacement occurs in the [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field).
The following query is equivalent to the previous one:
```logsql
_time:5m | replace ("secret-password", "***")
```
The number of replacements can be limited with `limit N` at the end of `replace`. For example, the following query replaces only the first `foo` substring with `bar`
at the [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) `baz`:
Performance tip: it is recommended using more specific [log filters](#filters) in order to reduce the number of log entries, which are passed to `replace`.
See [general performance tips](#performance-tips) for details.
`| replace_regexp ("regexp", "replacement") at field` [pipe](#pipes) replaces all the substrings matching the given `regexp` with the given `replacement`
in the given [`field`](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
The `regexp` must contain regular expression with [RE2 syntax](https://github.com/google/re2/wiki/Syntax).
The `replacement` may contain `$N` or `${N}` placeholders, which are substituted with the `N-th` capturing group in the `regexp`.
For example, the following query replaces all the substrings starting with `host-` and ending with `-foo` with the contents between `host-` and `-foo` in the [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field) for logs over the last 5 minutes:
```logsql
_time:5m | replace_regexp ("host-(.+?)-foo", "$1") at _msg
```
The `at _msg` part can be omitted if the replacement occurs in the [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field).
The following query is equivalent to the previous one:
The number of replacements can be limited with `limit N` at the end of `replace`. For example, the following query replaces only the first `password: ...` substring
ending with whitespace with empty substring at the [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) `baz`:
If the [`replace_regexp` pipe](#replace-pipe) mustn't be applied to every [log entry](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model),
By default logs are selected in arbitrary order because of performance reasons. If logs must be sorted, then `| sort by (field1, ..., fieldN)` [pipe](#pipes) can be used.
For example, the following query returns logs for the last 5 minutes sorted by [`_stream`](https://docs.victoriametrics.com/victorialogs/keyconcepts/#stream-fields)
and then by [`_time`](https://docs.victoriametrics.com/victorialogs/keyconcepts/#time-field):
```logsql
_time:5m | sort by (_stream, _time)
```
Add `desc` after the given log field in order to sort in reverse order of this field. For example, the following query sorts log fields in reverse order of `request_duration_seconds` field:
```logsql
_time:5m | sort by (request_duration_seconds desc)
```
The reverse order can be applied globally via `desc` keyword after `by(...)` clause:
the following query skips the first 10 logs with the biggest `request_duration` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model),
`| stats ...` pipe has the following basic format:
```logsql
... | stats
stats_func1(...) as result_name1,
...
stats_funcN(...) as result_nameN
```
Where `stats_func*` is any of the supported [stats function](#stats-pipe-functions), while `result_name*` is the name of the log field
to store the result of the corresponding stats function. The `as` keyword is optional.
For example, the following query calculates the following stats for logs over the last 5 minutes:
- the number of logs with the help of [`count` stats function](#count-stats);
- the number of unique [log streams](https://docs.victoriametrics.com/victorialogs/keyconcepts/#stream-fields) with the help of [`count_uniq` stats function](#count_uniq-stats):
It is allowed omitting the result name. In this case the result name equals to the string representation of the used [stats function](#stats-pipe-functions).
For example, the following query returns the same stats as the previous one, but gives uses `count()` and `count_uniq(_stream)` names for the returned fields:
The following LogsQL syntax can be used for calculating independent stats per group of log fields:
```logsql
... | stats by (field1, ..., fieldM)
stats_func1(...) as result_name1,
...
stats_funcN(...) as result_nameN
```
This calculates `stats_func*` per each `(field1, ..., fieldM)` group of [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
For example, the following query calculates the number of logs and unique ip addresses over the last 5 minutes,
grouped by `(host, path)` fields:
```logsql
_time:5m | stats by (host, path) count() logs_total, count_uniq(ip) ips_total
The following syntax can be used for calculating stats grouped by time buckets:
```logsql
... | stats by (_time:step)
stats_func1(...) as result_name1,
...
stats_funcN(...) as result_nameN
```
This calculates `stats_func*` per each `step` of [`_time`](https://docs.victoriametrics.com/victorialogs/keyconcepts/#time-field) field.
The `step` can have any [duration value](#duration-values). For example, the following LogsQL query returns per-minute number of logs and unique ip addresses
over the last 5 minutes:
```
_time:5m | stats by (_time:1m) count() logs_total, count_uniq(ip) ips_total
```
Additionally, the following `step` values are supported:
-`nanosecond` - equals to `1ns` [duration](#duration-values).
-`microsecond` - equals to `1µs` [duration](#duration-values).
-`millisecond` - equals to `1ms` [duration](#duration-values).
-`second` - equals to `1s` [duration](#duration-values).
-`minute` - equals to `1m` [duration](#duration-values).
-`day` - equals to `1d` [duration](#duration-values).
-`week` - equals to `1w` [duration](#duration-values).
-`month` - equals to one month. It properly takes into account the number of days per each month.
-`year` - equals to one year. It properly takes into account the number of days per each year.
#### Stats by time buckets with timezone offset
VictoriaLogs stores [`_time`](https://docs.victoriametrics.com/victorialogs/keyconcepts/#time-field) values as [Unix time](https://en.wikipedia.org/wiki/Unix_time)
in nanoseconds. This time corresponds to [UTC](https://en.wikipedia.org/wiki/Coordinated_Universal_Time) time zone. Sometimes it is needed calculating stats
grouped by days or weeks at non-UTC timezone. This is possible with the following syntax:
```logsql
... | stats by (_time:step offset timezone_offset) ...
```
For example, the following query calculates per-day number of logs over the last week, in `UTC+02:00` [time zone](https://en.wikipedia.org/wiki/Time_zone):
```logsql
_time:1w | stats by (_time:1d offset 2h) count() logs_total
```
#### Stats by field buckets
Every log field inside `| stats by (...)` can be bucketed in the same way at `_time` field in [this example](#stats-by-time-buckets).
Any [numeric value](#numeric-values) can be used as `step` value for the bucket. For example, the following query calculates
the number of requests for the last hour, bucketed by 10KB of `request_size_bytes` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model):
```logsql
_time:1h | stats by (request_size_bytes:10KB) count() requests
```
#### Stats by IPv4 buckets
Stats can be bucketed by [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) containing [IPv4 addresses](https://en.wikipedia.org/wiki/IP_address)
via the `ip_field_name:/network_mask` syntax inside `by(...)` clause. For example, the following query returns the number of log entries per `/24` subnetwork
extracted from the `ip` [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) during the last 5 minutes:
```logsql
_time:5m | stats by (ip:/24) count() requests_per_subnet
Sometimes it is needed to calculate stats on different subsets of matching logs. This can be done by inserting `if (<any_filters>)` condition
between [stats function](#stats-pipe-functions) and `result_name`, where `any_filter` can contain arbitrary [filters](#filters).
For example, the following query calculates individually the number of [logs messages](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field)
with `GET`, `POST` and `PUT` [words](#word), additionally to the total number of logs over the last 5 minutes:
`| top N by (field1, ..., fieldN)` [pipe](#pipes) returns top `N` sets for `(field1, ..., fieldN)` [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
with the maximum number of matching log entries.
For example, the following query returns top 7 [log streams](https://docs.victoriametrics.com/victorialogs/keyconcepts/#stream-fields)
with the maximum number of log entries over the last 5 minutes:
```logsql
_time:5m | top 7 by (_stream)
```
The `N` is optional. If it is skipped, then top 10 entries are returned. For example, the following query returns top 10 values
for `ip` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) seen in logs for the last 5 minutes:
```logsql
_time:5m | top by (ip)
```
The `by (...)` part in the `top` [pipe](#pipes) is optional. If it is skipped, then all the log fields are taken into account
when determining top field sets. This is useful when the field sets are already limited by other pipes such as [`fields` pipe](#fields-pipe).
For example, the following query is equivalent to the previous one:
Unique entries are stored in memory during query execution. Big number of unique selected entries may require a lot of memory.
Sometimes it is enough to return up to `N` unique entries. This can be done by adding `limit N` after `by (...)` clause.
This allows limiting memory usage. For example, the following query returns up to 100 unique `(host, path)` pairs for the logs over the last 5 minutes:
`| unpack_json from field_name` [pipe](#pipes) unpacks `{"k1":"v1", ..., "kN":"vN"}` JSON from the given input [`field_name`](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
into `k1`, ... `kN` output field names with the corresponding `v1`, ..., `vN` values. It overrides existing fields with names from the `k1`, ..., `kN` list. Other fields remain untouched.
Nested JSON is unpacked according to the rules defined [here](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
For example, the following query unpacks JSON fields from the [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field) across logs for the last 5 minutes:
The `from _msg` part can be omitted when JSON fields are unpacked from the [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field).
If only some fields must be extracted from JSON, then they can be enumerated inside `fields (...)`. For example, the following query unpacks only `foo` and `bar`
fields from JSON value stored in `my_json` [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model):
If it is needed to preserve the original non-empty field values, then add `keep_original_fields` to the end of `unpack_json ...`. For example,
the following query preserves the original non-empty values for `ip` and `host` fields instead of overwriting them with the unpacked values:
```logsql
_time:5m | unpack_json from foo fields (ip, host) keep_original_fields
```
Add `skip_empty_results` to the end of `unpack_json ...` if the original field values must be preserved when the corresponding unpacked values are empty.
For example, the following query preserves the original `ip` and `host` field values for empty unpacked values:
Performance tip: if you need extracting a single field from long JSON, it is faster to use [`extract` pipe](#extract-pipe). For example, the following query extracts `"ip"` field from JSON
If you want to make sure that the unpacked JSON fields do not clash with the existing fields, then specify common prefix for all the fields extracted from JSON,
by adding `result_prefix "prefix_name"` to `unpack_json`. For example, the following query adds `foo_` prefix for all the unpacked fields
form `foo`:
```logsql
_time:5m | unpack_json from foo result_prefix "foo_"
If the [`unpack_json` pipe](#unpack_json-pipe) mustn't be applied to every [log entry](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model),
The `<filters>` can contain arbitrary [filters](#filters). For example, the following query unpacks JSON fields from `foo` field only if `ip` field in the current log entry isn't set or empty:
with the corresponding `v1`, ..., `vN` values. It overrides existing fields with names from the `k1`, ..., `kN` list. Other fields remain untouched.
For example, the following query unpacks [logfmt](https://brandur.org/logfmt) fields from the [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field)
The `from _msg` part can be omitted when [logfmt](https://brandur.org/logfmt) fields are unpacked from the [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field).
If only some fields must be unpacked from logfmt, then they can be enumerated inside `fields (...)`. For example, the following query extracts only `foo` and `bar` fields
If it is needed to preserve the original non-empty field values, then add `keep_original_fields` to the end of `unpack_logfmt ...`. For example,
the following query preserves the original non-empty values for `ip` and `host` fields instead of overwriting them with the unpacked values:
```logsql
_time:5m | unpack_logfmt from foo fields (ip, host) keep_original_fields
```
Add `skip_empty_results` to the end of `unpack_logfmt ...` if the original field values must be preserved when the corresponding unpacked values are empty.
For example, the following query preserves the original `ip` and `host` field values for empty unpacked values:
Performance tip: if you need extracting a single field from long [logfmt](https://brandur.org/logfmt) line, it is faster to use [`extract` pipe](#extract-pipe).
For example, the following query extracts `"ip"` field from [logfmt](https://brandur.org/logfmt) line stored
in [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field):
If you want to make sure that the unpacked [logfmt](https://brandur.org/logfmt) fields do not clash with the existing fields, then specify common prefix for all the fields extracted from logfmt,
If the [`unpack_logfmt` pipe](#unpack_logfmt-pipe) mustn't be applied to every [log entry](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model),
-`timestamp` - timestamp in [ISO8601 format](https://en.wikipedia.org/wiki/ISO_8601). The `MMM DD hh:mm:ss` timestamp in [RFC3164](https://datatracker.ietf.org/doc/html/rfc3164)
is automatically converted into [ISO8601 format](https://en.wikipedia.org/wiki/ISO_8601) by assuming that the timestamp belongs to the last 12 months.
The `[STRUCTURED-DATA]` is parsed into fields with the `SD-ID` name and `param1="value1" ... paramN="valueN"` value
according to [the specification](https://datatracker.ietf.org/doc/html/rfc5424#section-6.3). The value then can be parsed to separate fields with [`unpack_logfmt` pipe](#unpack_logfmt-pipe).
For example, the following query unpacks [syslog](https://en.wikipedia.org/wiki/Syslog) message from the [`_msg` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field)
By default timestamps in [RFC3164 format](https://datatracker.ietf.org/doc/html/rfc3164) are converted to local timezone. It is possible to change the timezone
offset via `offset` option. For example, the following query adds 5 hours and 30 minutes to unpacked `rfc3164` timestamps:
If the [`unpack_syslog` pipe](#unpack_syslog-pipe) musn't be applied to every [log entry](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model),
then add `if (<filters>)` after `unpack_syslog`.
The `<filters>` can contain arbitrary [filters](#filters). For example, the following query unpacks syslog message fields from `foo` field
only if `hostname` field in the current log entry isn't set or empty:
```logsql
_time:5m | unpack_syslog if (hostname:"") from foo
`| unroll by (field1, ..., fieldN)` [pipe](#pipes) can be used for unrolling JSON arrays from `field1`, `fieldN`
[log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) into separate rows.
For example, the following query unrolls `timestamp` and `value` [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) from logs for the last 5 minutes:
If the [`unroll` pipe](#unpack_logfmt-pipe) mustn't be applied to every [log entry](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model),
The `<filters>` can contain arbitrary [filters](#filters). For example, the following query unrolls `value` field only if `value_type` field equals to `json_array`:
```logsql
_time:5m | unroll if (value_type:="json_array") (value)
- [`avg`](#avg-stats) returns the average value over the given numeric [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
- [`count`](#count-stats) returns the number of log entries.
- [`count_empty`](#count_empty-stats) returns the number logs with empty [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
- [`count_uniq`](#count_uniq-stats) returns the number of unique non-empty values for the given [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
- [`max`](#max-stats) returns the maximum value over the given numeric [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
- [`median`](#median-stats) returns the [median](https://en.wikipedia.org/wiki/Median) value over the given numeric [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
- [`min`](#min-stats) returns the minumum value over the given numeric [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
- [`quantile`](#quantile-stats) returns the given quantile for the given numeric [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
- [`row_any`](#row_any-stats) returns a sample [log entry](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) per each selected [stats group](#stats-by-fields).
- [`row_max`](#row_max-stats) returns the [log entry](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) with the minimum value at the given field.
- [`row_min`](#row_min-stats) returns the [log entry](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) with the maximum value at the given field.
- [`sum`](#sum-stats) returns the sum for the given numeric [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
- [`sum_len`](#sum_len-stats) returns the sum of lengths for the given [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
- [`uniq_values`](#uniq_values-stats) returns unique non-empty values for the given [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
- [`values`](#values-stats) returns all the values for the given [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
all the mentioned [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
Non-numeric values are ignored.
For example, the following query returns the average value for the `duration` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
For example, the following query returns the number of logs over the last 5 minutes:
```logsql
_time:5m | stats count() logs
```
It is possible calculating the number of logs with non-empty values for some [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
with the `count(fieldName)` syntax. For example, the following query returns the number of logs with non-empty `username` field over the last 5 minutes:
If multiple fields are enumerated inside `count()`, then it counts the number of logs with at least a single non-empty field mentioned inside `count()`.
For example, the following query returns the number of logs with non-empty `username` or `password` [fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
`count_empty(field1, ..., fieldN)` [stats pipe function](#stats-pipe-functions) calculates the number of logs with empty `(field1, ..., fieldN)` tuples.
For example, the following query calculates the number of logs with empty `username` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
For example, the following query returns the number of unique non-empty values for `ip` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
over the last 5 minutes:
```logsql
_time:5m | stats count_uniq(ip) ips
```
The following query returns the number of unique `(host, path)` pairs for the corresponding [fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
all the mentioned [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
For example, the following query returns the maximum value for the `duration` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
all the mentioned [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
For example, the following query returns the minimum value for the `duration` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
for the given [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model). The `phi` must be in the range `0 ... 1`, where `0` means `0th` percentile,
while `1` means `100th` percentile.
For example, the following query calculates `50th`, `90th` and `99th` percentiles for the `request_duration_seconds` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
(aka sample) per each selected [stats group](#stats-by-fields). Log entry is returned as JSON-encoded dictionary with all the fields from the original log.
For example, the following query returns a sample log entry per each [`_stream`](https://docs.victoriametrics.com/victorialogs/keyconcepts/#stream-fields)
across logs for the last 5 minutes:
```logsql
_time:5m | stats by (_stream) row_any() as sample_row
```
Fields from the returned values can be decoded with [`unpack_json`](#unpack_json-pipe) or [`extract`](#extract-pipe) pipes.
If only the specific fields are needed, then they can be enumerated inside `row_any(...)`.
with the maximum value for the given `field`. Log entry is returned as JSON-encoded dictionary with all the fields from the original log.
For example, the following query returns log entry with the maximum value for the `duration` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
across logs for the last 5 minutes:
```logsql
_time:5m | stats row_max(duration) as log_with_max_duration
```
Fields from the returned values can be decoded with [`unpack_json`](#unpack_json-pipe) or [`extract`](#extract-pipe) pipes.
If only the specific fields are needed from the returned log entry, then they can be enumerated inside `row_max(...)`.
For example, the following query returns only `_time`, `path` and `duration` fields from the log entry with the maximum `duration` over the last 5 minutes:
```logsql
_time:5m | stats row_max(duration, _time, path, duration) as time_and_path_with_max_duration
with the minimum value for the given `field`. Log entry is returned as JSON-encoded dictionary with all the fields from the original log.
For example, the following query returns log entry with the minimum value for the `duration` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
across logs for the last 5 minutes:
```logsql
_time:5m | stats row_min(duration) as log_with_min_duration
```
Fields from the returned values can be decoded with [`unpack_json`](#unpack_json-pipe) or [`extract`](#extract-pipe) pipes.
If only the specific fields are needed from the returned log entry, then they can be enumerated inside `row_max(...)`.
For example, the following query returns only `_time`, `path` and `duration` fields from the log entry with the minimum `duration` over the last 5 minutes:
```logsql
_time:5m | stats row_min(duration, _time, path, duration) as time_and_path_with_min_duration
all the mentioned [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
For example, the following query returns the sum of numeric values for the `duration` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
for the given [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
For example, the following query returns the sum of lengths of [`_msg` fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#message-field)
For example, the following query returns unique non-empty values for the `ip` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
Every unique value is stored in memory during query execution. Big number of unique values may require a lot of memory. Sometimes it is enough to return
only a subset of unique values. In this case add `limit N` after `uniq_values(...)` in order to limit the number of returned unique values to `N`,
while limiting the maximum memory usage.
For example, the following query returns up to `100` unique values for the `ip` [field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model)
- Extracting arbitrary text from [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) according to the provided pattern.
- Unpacking JSON fields from [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model). See [these docs](#unpack_json-pipe).
- Unpacking [logfmt](https://brandur.org/logfmt) fields from [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model). See [these docs](#unpack_logfmt-pipe).
- Unpacking [Syslog](https://en.wikipedia.org/wiki/Syslog) messages from [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model). See [these docs](#unpack_syslog-pipe).
- Creating a new field from existing [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) according to the provided format. See [`format` pipe](#format-pipe).
- Replacing substrings in the given [log field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model).
See [`replace` pipe](#replace-pipe) and [`replace_regexp` pipe](#replace_regexp-pipe) docs.
- Creating a new field according to math calculations over existing [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model). See [`math` pipe](#math-pipe).
- [`fields`](#fields-pipe) and [`delete`](#delete-pipe) pipes allow limiting the set of [log fields](https://docs.victoriametrics.com/victorialogs/keyconcepts/#data-model) to return.
