Merge #145932

145932: sql: remove old hyperloglog version r=yuzefovich a=yuzefovich

This was only needed for compatibility with 24.3 nodes to ensure that table stats collection worked in mixed-version state. We no longer need compatibility with that version. Additionally remove no longer used sketch type field from the processor spec.

Epic: None
Release note: None

Co-authored-by: Yahor Yuzefovich <[email protected]>

Author: craig[bot]

DEPS.bzl

@@ -793,16 +793,6 @@ def go_deps():
             "https://storage.googleapis.com/cockroach-godeps/gomod/github.com/axiomhq/hyperloglog/com_github_axiomhq_hyperloglog-v0.2.5.zip",
         ],
     )
-    go_repository(
-        name = "com_github_axiomhq_hyperloglog_000",
-        build_file_proto_mode = "disable_global",
-        importpath = "github.com/axiomhq/hyperloglog/000",
-        sha256 = "812834322ee2ca50dc36f91f9ac3f2cde4631af2f9c330b1271c78b46024a540",
-        strip_prefix = "github.com/axiomhq/[email protected]",
-        urls = [
-            "https://storage.googleapis.com/cockroach-godeps/gomod/github.com/axiomhq/hyperloglog/com_github_axiomhq_hyperloglog-v0.0.0-20181223111420-4b99d0c2c99e.zip",
-        ],
-    )
     go_repository(
         name = "com_github_aymanbagabas_go_osc52",
         build_file_proto_mode = "disable_global",

build/bazelutil/distdir_files.bzl

@@ -286,7 +286,6 @@ DISTDIR_FILES = {
     "https://storage.googleapis.com/cockroach-godeps/gomod/github.com/aws/aws-sdk-go-v2/service/sts/com_github_aws_aws_sdk_go_v2_service_sts-v1.33.17.zip": "87aca25fafd483a1eac29c5baaab05ad485422a9aa1ccc5db0d39733c2d71cd2",
     "https://storage.googleapis.com/cockroach-godeps/gomod/github.com/aws/aws-sdk-go/com_github_aws_aws_sdk_go-v1.40.37.zip": "c0c481d28af88f621fb3fdeacc1e5d32f69a1bb83d0ee959f95ce89e4e2d0494",
     "https://storage.googleapis.com/cockroach-godeps/gomod/github.com/aws/smithy-go/com_github_aws_smithy_go-v1.22.3.zip": "572df48de9133d57f45909d3067b2053b97230268c2d28e4e44ea9644009ef11",
-    "https://storage.googleapis.com/cockroach-godeps/gomod/github.com/axiomhq/hyperloglog/com_github_axiomhq_hyperloglog-v0.0.0-20181223111420-4b99d0c2c99e.zip": "812834322ee2ca50dc36f91f9ac3f2cde4631af2f9c330b1271c78b46024a540",
     "https://storage.googleapis.com/cockroach-godeps/gomod/github.com/axiomhq/hyperloglog/com_github_axiomhq_hyperloglog-v0.2.5.zip": "6125b12664bb5dd8614e82f0fe7528242dcb11649e1d7e051aabf3da471e14e1",
     "https://storage.googleapis.com/cockroach-godeps/gomod/github.com/aymanbagabas/go-osc52/com_github_aymanbagabas_go_osc52-v1.0.3.zip": "138e75a51599c2a8e4afe2bd6acdeaddbb73eb9ec796dfa2f577b16201660d9e",
     "https://storage.googleapis.com/cockroach-godeps/gomod/github.com/aymerick/douceur/com_github_aymerick_douceur-v0.2.0.zip": "dcbf69760cc1a8b32384495438e1086e4c3d669b2ebc0debd92e1865ffd6be60",

go.mod

@@ -120,7 +120,6 @@ require (
 	github.com/aws/aws-sdk-go-v2/service/secretsmanager v1.35.1
 	github.com/aws/smithy-go v1.22.3
 	github.com/axiomhq/hyperloglog v0.2.5
-	github.com/axiomhq/hyperloglog/000 v0.0.0-20181223111420-4b99d0c2c99e
 	github.com/bazelbuild/rules_go v0.26.0
 	github.com/biogo/store v0.0.0-20160505134755-913427a1d5e8
 	github.com/blevesearch/snowballstem v0.9.0
@@ -500,10 +499,6 @@ replace golang.org/x/time => github.com/cockroachdb/x-time v0.3.1-0.202305251236
 
 replace github.com/gogo/protobuf => github.com/cockroachdb/gogoproto v1.3.3-0.20241216150617-2358cdb156a1
 
-// TODO(yuzefovich): remove this version once compatibility with 24.3 is no
-// longer needed.
-replace github.com/axiomhq/hyperloglog/000 => github.com/axiomhq/hyperloglog v0.0.0-20181223111420-4b99d0c2c99e
-
 // Note: This forked dependency adds a commit that opens up some
 // private APIs to enable us to make some perf improvements to
 // histogram updates in particular.

go.sum

@@ -447,8 +447,6 @@ github.com/aws/aws-sdk-go-v2/service/sts v1.33.17/go.mod h1:cQnB8CUnxbMU82JvlqjK
 github.com/aws/smithy-go v1.13.5/go.mod h1:Tg+OJXh4MB2R/uN61Ko2f6hTZwB/ZYGOtib8J3gBHzA=
 github.com/aws/smithy-go v1.22.3 h1:Z//5NuZCSW6R4PhQ93hShNbyBbn8BWCmCVCt+Q8Io5k=
 github.com/aws/smithy-go v1.22.3/go.mod h1:t1ufH5HMublsJYulve2RKmHDC15xu1f26kHCp/HgceI=
-github.com/axiomhq/hyperloglog v0.0.0-20181223111420-4b99d0c2c99e h1:190ugM9MsyFauTkR/UqcHG/mn5nmFe6SvHJqEHIrtrA=
-github.com/axiomhq/hyperloglog v0.0.0-20181223111420-4b99d0c2c99e/go.mod h1:IOXAcuKIFq/mDyuQ4wyJuJ79XLMsmLM+5RdQ+vWrL7o=
 github.com/axiomhq/hyperloglog v0.2.5 h1:Hefy3i8nAs8zAI/tDp+wE7N+Ltr8JnwiW3875pvl0N8=
 github.com/axiomhq/hyperloglog v0.2.5/go.mod h1:DLUK9yIzpU5B6YFLjxTIcbHu1g4Y1WQb1m5RH3radaM=
 github.com/aymanbagabas/go-osc52 v1.0.3 h1:DTwqENW7X9arYimJrPeGZcV0ln14sGMt3pHZspWD+Mg=

pkg/sql/distsql_plan_stats.go

@@ -410,7 +410,6 @@ func (dsp *DistSQLPlanner) createPartialStatsPlan(
 
 	sampledColumnIDs := make([]descpb.ColumnID, len(scan.catalogCols))
 	spec := execinfrapb.SketchSpec{
-		SketchType:          execinfrapb.SketchType_HLL_PLUS_PLUS_V1,
 		GenerateHistogram:   reqStat.histogram,
 		HistogramMaxBuckets: reqStat.histogramMaxBuckets,
 		Columns:             make([]uint32, len(reqStat.columns)),
@@ -656,7 +655,6 @@ func (dsp *DistSQLPlanner) createStatsPlan(
 	sampledColumnIDs := make([]descpb.ColumnID, len(requestedCols))
 	for _, s := range reqStats {
 		spec := execinfrapb.SketchSpec{
-			SketchType:          execinfrapb.SketchType_HLL_PLUS_PLUS_V1,
 			GenerateHistogram:   s.histogram,
 			HistogramMaxBuckets: s.histogramMaxBuckets,
 			Columns:             make([]uint32, len(s.columns)),

pkg/sql/execinfrapb/processors_table_stats.proto

@@ -17,18 +17,9 @@ option go_package = "github.com/cockroachdb/cockroach/pkg/sql/execinfrapb";
 import "sql/catalog/descpb/structured.proto";
 import "gogoproto/gogo.proto";
 
-// TODO(yuzefovich): this can be removed once compatibility with 24.3 is no
-// longer needed.
-enum SketchType {
-  // This is the github.com/axiomhq/hyperloglog binary format (as of commit
-  // 730eea1) for a sketch with precision 14. Values are encoded using their key
-  // encoding, except integers which are encoded in 8 bytes (little-endian).
-  HLL_PLUS_PLUS_V1 = 0;
-}
-
 // SketchSpec contains the specification for a generated statistic.
 message SketchSpec {
-  optional SketchType sketch_type = 1 [(gogoproto.nullable) = false];
+  reserved 1;
 
   // Each value is an index identifying a column in the input stream.
   // TODO(radu): currently only one column is supported.

pkg/sql/rowexec/BUILD.bazel

@@ -115,7 +115,6 @@ go_library(
         "//pkg/util/tracing/tracingpb",
         "//pkg/util/vector",
         "@com_github_axiomhq_hyperloglog//:hyperloglog",
-        "@com_github_axiomhq_hyperloglog_000//:000",
         "@com_github_cockroachdb_errors//:errors",
         "@com_github_cockroachdb_logtags//:logtags",
         "@com_github_cockroachdb_redact//:redact",

pkg/sql/rowexec/sample_aggregator.go

@@ -10,14 +10,12 @@ import (
 	"math"
 	"time"
 
-	hllNew "github.com/axiomhq/hyperloglog"
-	hllOld "github.com/axiomhq/hyperloglog/000"
+	"github.com/axiomhq/hyperloglog"
 	"github.com/cockroachdb/cockroach/pkg/jobs"
 	"github.com/cockroachdb/cockroach/pkg/server/telemetry"
 	"github.com/cockroachdb/cockroach/pkg/sql/catalog/descpb"
 	"github.com/cockroachdb/cockroach/pkg/sql/execinfra"
 	"github.com/cockroachdb/cockroach/pkg/sql/execinfrapb"
-	"github.com/cockroachdb/cockroach/pkg/sql/execversion"
 	"github.com/cockroachdb/cockroach/pkg/sql/isql"
 	"github.com/cockroachdb/cockroach/pkg/sql/parser"
 	"github.com/cockroachdb/cockroach/pkg/sql/pgwire/pgcode"
@@ -103,7 +101,6 @@ func newSampleAggregator(
 			return nil, errors.Errorf("histograms require one column")
 		}
 	}
-	useNewHLL := execversion.FromContext(ctx) >= execversion.V25_1
 
 	// Limit the memory use by creating a child monitor with a hard limit.
 	// The processor will disable histogram collection if this limit is not
@@ -144,14 +141,10 @@ func newSampleAggregator(
 	for i := range spec.Sketches {
 		s.sketches[i] = sketchInfo{
 			spec:     spec.Sketches[i],
+			sketch:   hyperloglog.New14(),
 			numNulls: 0,
 			numRows:  0,
 		}
-		if useNewHLL {
-			s.sketches[i].sketchNew = hllNew.New14()
-		} else {
-			s.sketches[i].sketchOld = hllOld.New14()
-		}
 		if spec.Sketches[i].GenerateHistogram {
 			sampleCols.Add(int(spec.Sketches[i].Columns[0]))
 		}
@@ -173,14 +166,10 @@ func newSampleAggregator(
 		s.invSr[col] = &sr
 		s.invSketch[col] = &sketchInfo{
 			spec:     spec.InvertedSketches[i],
+			sketch:   hyperloglog.New14(),
 			numNulls: 0,
 			numRows:  0,
 		}
-		if useNewHLL {
-			s.invSketch[col].sketchNew = hllNew.New14()
-		} else {
-			s.invSketch[col].sketchOld = hllOld.New14()
-		}
 	}
 
 	if err := s.Init(
@@ -372,6 +361,8 @@ func (s *sampleAggregator) mainLoop(
 func (s *sampleAggregator) processSketchRow(
 	sketch *sketchInfo, row rowenc.EncDatumRow, da *tree.DatumAlloc,
 ) error {
+	var tmpSketch hyperloglog.Sketch
+
 	numRows, err := row[s.numRowsCol].GetInt()
 	if err != nil {
 		return err
@@ -398,22 +389,11 @@ func (s *sampleAggregator) processSketchRow(
 	if d == tree.DNull {
 		return errors.AssertionFailedf("NULL sketch data")
 	}
-	if sketch.sketchNew != nil {
-		var tmpSketch hllNew.Sketch
-		if err := tmpSketch.UnmarshalBinary([]byte(*d.(*tree.DBytes))); err != nil {
-			return err
-		}
-		if err := sketch.sketchNew.Merge(&tmpSketch); err != nil {
-			return errors.NewAssertionErrorWithWrappedErrf(err, "merging sketch data")
-		}
-	} else {
-		var tmpSketch hllOld.Sketch
-		if err := tmpSketch.UnmarshalBinary([]byte(*d.(*tree.DBytes))); err != nil {
-			return err
-		}
-		if err := sketch.sketchOld.Merge(&tmpSketch); err != nil {
-			return errors.NewAssertionErrorWithWrappedErrf(err, "merging sketch data")
-		}
+	if err := tmpSketch.UnmarshalBinary([]byte(*d.(*tree.DBytes))); err != nil {
+		return err
+	}
+	if err := sketch.sketch.Merge(&tmpSketch); err != nil {
+		return errors.NewAssertionErrorWithWrappedErrf(err, "merging sketch data")
 	}
 	return nil
 }
@@ -631,12 +611,7 @@ func (s *sampleAggregator) getAvgSize(si *sketchInfo) int64 {
 // getDistinctCount returns the number of distinct values in the given sketch,
 // optionally including null values.
 func (s *sampleAggregator) getDistinctCount(si *sketchInfo, includeNulls bool) int64 {
-	var distinctCount int64
-	if si.sketchNew != nil {
-		distinctCount = int64(si.sketchNew.Estimate())
-	} else {
-		distinctCount = int64(si.sketchOld.Estimate())
-	}
+	distinctCount := int64(si.sketch.Estimate())
 	if si.numNulls > 0 && !includeNulls {
 		// Nulls are included in the estimate, so reduce the count by 1 if nulls are
 		// not requested.

pkg/sql/rowexec/sampler.go

@@ -11,8 +11,7 @@ import (
 	"math/rand"
 	"time"
 
-	hllNew "github.com/axiomhq/hyperloglog"
-	hllOld "github.com/axiomhq/hyperloglog/000"
+	"github.com/axiomhq/hyperloglog"
 	"github.com/cockroachdb/cockroach/pkg/server/telemetry"
 	"github.com/cockroachdb/cockroach/pkg/sql/catalog/catenumpb"
 	"github.com/cockroachdb/cockroach/pkg/sql/execinfra"
@@ -36,10 +35,8 @@ import (
 
 // sketchInfo contains the specification and run-time state for each sketch.
 type sketchInfo struct {
-	spec execinfrapb.SketchSpec
-	// Exactly one of sketchOld and sketchNew will be set.
-	sketchOld            *hllOld.Sketch
-	sketchNew            *hllNew.Sketch
+	spec                 execinfrapb.SketchSpec
+	sketch               *hyperloglog.Sketch
 	numNulls             int64
 	numRows              int64
 	size                 int64
@@ -103,7 +100,6 @@ func newSamplerProcessor(
 	input execinfra.RowSource,
 	post *execinfrapb.PostProcessSpec,
 ) (*samplerProcessor, error) {
-	useNewHLL := execversion.FromContext(ctx) >= execversion.V25_1
 	legacyFingerprinting := execversion.FromContext(ctx) < execversion.V25_2
 
 	// Limit the memory use by creating a child monitor with a hard limit.
@@ -133,15 +129,11 @@ func newSamplerProcessor(
 	for i := range spec.Sketches {
 		s.sketches[i] = sketchInfo{
 			spec:                 spec.Sketches[i],
+			sketch:               hyperloglog.New14(),
 			numNulls:             0,
 			numRows:              0,
 			legacyFingerprinting: legacyFingerprinting,
 		}
-		if useNewHLL {
-			s.sketches[i].sketchNew = hllNew.New14()
-		} else {
-			s.sketches[i].sketchOld = hllOld.New14()
-		}
 		if spec.Sketches[i].GenerateHistogram {
 			sampleCols.Add(int(spec.Sketches[i].Columns[0]))
 		}
@@ -160,14 +152,10 @@ func newSamplerProcessor(
 		sketchSpec.Columns = []uint32{0}
 		s.invSketch[col] = &sketchInfo{
 			spec:     sketchSpec,
+			sketch:   hyperloglog.New14(),
 			numNulls: 0,
 			numRows:  0,
 		}
-		if useNewHLL {
-			s.invSketch[col].sketchNew = hllNew.New14()
-		} else {
-			s.invSketch[col].sketchOld = hllOld.New14()
-		}
 	}
 
 	s.sr.Init(int(spec.SampleSize), int(spec.MinSampleSize), inTypes, &s.memAcc, sampleCols)
@@ -437,12 +425,7 @@ func (s *samplerProcessor) emitSketchRow(
 	outRow[s.numRowsCol] = rowenc.EncDatum{Datum: tree.NewDInt(tree.DInt(si.numRows))}
 	outRow[s.numNullsCol] = rowenc.EncDatum{Datum: tree.NewDInt(tree.DInt(si.numNulls))}
 	outRow[s.sizeCol] = rowenc.EncDatum{Datum: tree.NewDInt(tree.DInt(si.size))}
-	var data []byte
-	if si.sketchNew != nil {
-		data, err = si.sketchNew.MarshalBinary()
-	} else {
-		data, err = si.sketchOld.MarshalBinary()
-	}
+	data, err := si.sketch.MarshalBinary()
 	if err != nil {
 		return false, err
 	}
@@ -603,11 +586,7 @@ func (s *sketchInfo) addRow(
 	if allNulls {
 		s.numNulls++
 	}
-	if s.sketchNew != nil {
-		s.sketchNew.Insert(*buf)
-	} else {
-		s.sketchOld.Insert(*buf)
-	}
+	s.sketch.Insert(*buf)
 	return nil
 }
 
@@ -678,11 +657,7 @@ func (s *sketchInfo) addRowLegacy(
 		// be uniformly distributed in the 2^64 range). Experiments (on tpcc
 		// order_line) with simplistic functions yielded bad results.
 		binary.LittleEndian.PutUint64(*buf, uint64(val))
-		if s.sketchNew != nil {
-			s.sketchNew.Insert(*buf)
-		} else {
-			s.sketchOld.Insert(*buf)
-		}
+		s.sketch.Insert(*buf)
 		return nil
 	}
 	isNull := true
@@ -715,10 +690,6 @@ func (s *sketchInfo) addRowLegacy(
 	if isNull {
 		s.numNulls++
 	}
-	if s.sketchNew != nil {
-		s.sketchNew.Insert(*buf)
-	} else {
-		s.sketchOld.Insert(*buf)
-	}
+	s.sketch.Insert(*buf)
 	return nil
 }