From d50273d6c0c103da2a457232a7410a4e674893ba Mon Sep 17 00:00:00 2001
From: Aliaksandr Valialkin <valyala@victoriametrics.com>
Date: Wed, 15 May 2024 22:31:21 +0200
Subject: [PATCH] wip

---
 docs/VictoriaLogs/LogsQL.md   | 18 ++++++++++++++++++
 lib/logstorage/parser_test.go |  3 +++
 lib/logstorage/pipe_sort.go   |  6 ++++--
 lib/logstorage/pipe_stats.go  |  6 ++++--
 lib/logstorage/pipe_uniq.go   |  6 ++++--
 5 files changed, 33 insertions(+), 6 deletions(-)

diff --git a/docs/VictoriaLogs/LogsQL.md b/docs/VictoriaLogs/LogsQL.md
index 4418248ed..700cb130b 100644
--- a/docs/VictoriaLogs/LogsQL.md
+++ b/docs/VictoriaLogs/LogsQL.md
@@ -1200,6 +1200,12 @@ The reverse order can be applied globally via `desc` keyword after `by(...)` cla
 _time:5m | sort by (foo, bar) desc
 ```
 
+The `by` keyword can be skipped in `sort ...` pipe. For example, the following query is equivalent to the previous one:
+
+```logsql
+_time:5m | sort (foo, bar) desc
+```
+
 Sorting of big number of logs can consume a lot of CPU time and memory. Sometimes it is enough to return the first `N` entries with the biggest
 or the smallest values. This can be done by adding `limit N` to the end of `sort ...` pipe.
 Such a query consumes lower amounts of memory when sorting big number of logs, since it keeps in memory only `N` log entries.
@@ -1257,6 +1263,12 @@ This allows limiting memory usage. For example, the following query returns up t
 _time:5m | uniq by (host, path) limit 100
 ```
 
+The `by` keyword can be skipped in `uniq ...` pipe. For example, the following query is equivalent to the previous one:
+
+```logsql
+_time:5m | uniq (host, path) limit 100
+```
+
 See also:
 
 - [`uniq_values` stats function](#uniq_values-stats)
@@ -1316,6 +1328,12 @@ grouped by `(host, path)` fields:
 _time:5m | stats by (host, path) count() logs_total, count_uniq(ip) ips_total
 ```
 
+The `by` keyword can be skipped in `stats ...` pipe. For example, the following query is equvalent to the previous one:
+
+```logsql
+_time:5m | stats (host, path) count() logs_total, count_uniq(ip) ips_total
+```
+
 #### Stats by time buckets
 
 The following syntax can be used for calculating stats grouped by time buckets:
diff --git a/lib/logstorage/parser_test.go b/lib/logstorage/parser_test.go
index 1f0289e58..2a021908a 100644
--- a/lib/logstorage/parser_test.go
+++ b/lib/logstorage/parser_test.go
@@ -966,6 +966,7 @@ func TestParseQuerySuccess(t *testing.T) {
 	f(`* | stats by (_time:week) count() foo`, `* | stats by (_time:week) count(*) as foo`)
 	f(`* | stats by (_time:month) count() foo`, `* | stats by (_time:month) count(*) as foo`)
 	f(`* | stats by (_time:year offset 6.5h) count() foo`, `* | stats by (_time:year offset 6.5h) count(*) as foo`)
+	f(`* | stats (_time:year offset 6.5h) count() foo`, `* | stats by (_time:year offset 6.5h) count(*) as foo`)
 
 	// sort pipe
 	f(`* | sort`, `* | sort`)
@@ -983,6 +984,7 @@ func TestParseQuerySuccess(t *testing.T) {
 	f(`* | sort by (foo desc, bar) desc limit 10`, `* | sort by (foo desc, bar) desc limit 10`)
 	f(`* | sort by (foo desc, bar) desc OFFSET 30 limit 10`, `* | sort by (foo desc, bar) desc offset 30 limit 10`)
 	f(`* | sort by (foo desc, bar) desc limit 10 OFFSET 30`, `* | sort by (foo desc, bar) desc offset 30 limit 10`)
+	f(`* | sort (foo desc, bar) desc limit 10 OFFSET 30`, `* | sort by (foo desc, bar) desc offset 30 limit 10`)
 
 	// uniq pipe
 	f(`* | uniq`, `* | uniq`)
@@ -991,6 +993,7 @@ func TestParseQuerySuccess(t *testing.T) {
 	f(`* | uniq by(foo,*,bar)`, `* | uniq`)
 	f(`* | uniq by(f1,f2)`, `* | uniq by (f1, f2)`)
 	f(`* | uniq by(f1,f2) limit 10`, `* | uniq by (f1, f2) limit 10`)
+	f(`* | uniq (f1,f2) limit 10`, `* | uniq by (f1, f2) limit 10`)
 	f(`* | uniq limit 10`, `* | uniq limit 10`)
 
 	// multiple different pipes
diff --git a/lib/logstorage/pipe_sort.go b/lib/logstorage/pipe_sort.go
index 55b74b5dd..f1bcc522a 100644
--- a/lib/logstorage/pipe_sort.go
+++ b/lib/logstorage/pipe_sort.go
@@ -689,8 +689,10 @@ func parsePipeSort(lex *lexer) (*pipeSort, error) {
 	lex.nextToken()
 
 	var ps pipeSort
-	if lex.isKeyword("by") {
-		lex.nextToken()
+	if lex.isKeyword("by", "(") {
+		if lex.isKeyword("by") {
+			lex.nextToken()
+		}
 		bfs, err := parseBySortFields(lex)
 		if err != nil {
 			return nil, fmt.Errorf("cannot parse 'by' clause: %w", err)
diff --git a/lib/logstorage/pipe_stats.go b/lib/logstorage/pipe_stats.go
index efb5a9cfd..62b3def5b 100644
--- a/lib/logstorage/pipe_stats.go
+++ b/lib/logstorage/pipe_stats.go
@@ -443,8 +443,10 @@ func parsePipeStats(lex *lexer) (*pipeStats, error) {
 	lex.nextToken()
 
 	var ps pipeStats
-	if lex.isKeyword("by") {
-		lex.nextToken()
+	if lex.isKeyword("by", "(") {
+		if lex.isKeyword("by") {
+			lex.nextToken()
+		}
 		bfs, err := parseByStatsFields(lex)
 		if err != nil {
 			return nil, fmt.Errorf("cannot parse 'by' clause: %w", err)
diff --git a/lib/logstorage/pipe_uniq.go b/lib/logstorage/pipe_uniq.go
index 3b72aaac2..3b2fcc66d 100644
--- a/lib/logstorage/pipe_uniq.go
+++ b/lib/logstorage/pipe_uniq.go
@@ -360,8 +360,10 @@ func parsePipeUniq(lex *lexer) (*pipeUniq, error) {
 	lex.nextToken()
 
 	var pu pipeUniq
-	if lex.isKeyword("by") {
-		lex.nextToken()
+	if lex.isKeyword("by", "(") {
+		if lex.isKeyword("by") {
+			lex.nextToken()
+		}
 		bfs, err := parseFieldNamesInParens(lex)
 		if err != nil {
 			return nil, fmt.Errorf("cannot parse 'by' clause: %w", err)