github.com/cockroachdb/cockroach@v20.2.0-alpha.1+incompatible/pkg/sql/colexec/hash_utils_test.go (about)

     1  // Copyright 2020 The Cockroach Authors.
     2  //
     3  // Use of this software is governed by the Business Source License
     4  // included in the file licenses/BSL.txt.
     5  //
     6  // As of the Change Date specified in that file, in accordance with
     7  // the Business Source License, use of this software will be governed
     8  // by the Apache License, Version 2.0, included in the file
     9  // licenses/APL.txt.
    10  
    11  package colexec
    12  
    13  import (
    14  	"context"
    15  	"fmt"
    16  	"testing"
    17  
    18  	"github.com/cockroachdb/cockroach/pkg/col/coldata"
    19  	"github.com/cockroachdb/cockroach/pkg/sql/sqlbase"
    20  	"github.com/cockroachdb/cockroach/pkg/sql/types"
    21  	"github.com/cockroachdb/cockroach/pkg/util/leaktest"
    22  )
    23  
    24  // TestHashFunctionFamily verifies the assumption that our vectorized hashing
    25  // function (the combination of initHash, rehash, and finalizeHash) actually
    26  // defines a function family and that changing the initial hash value is
    27  // sufficient to get a "different" hash function.
    28  func TestHashFunctionFamily(t *testing.T) {
    29  	defer leaktest.AfterTest(t)()
    30  
    31  	ctx := context.Background()
    32  	bucketsA, bucketsB := make([]uint64, coldata.BatchSize()), make([]uint64, coldata.BatchSize())
    33  	nKeys := coldata.BatchSize()
    34  	keyTypes := []*types.T{types.Int}
    35  	keys := []coldata.Vec{testAllocator.NewMemColumn(keyTypes[0], coldata.BatchSize())}
    36  	for i := int64(0); i < int64(coldata.BatchSize()); i++ {
    37  		keys[0].Int64()[i] = i
    38  	}
    39  	numBuckets := uint64(16)
    40  	var (
    41  		cancelChecker     CancelChecker
    42  		overloadHelperVar overloadHelper
    43  		datumAlloc        sqlbase.DatumAlloc
    44  	)
    45  
    46  	for initHashValue, buckets := range [][]uint64{bucketsA, bucketsB} {
    47  		// We need +1 here because 0 is not a valid initial hash value.
    48  		initHash(buckets, nKeys, uint64(initHashValue+1))
    49  		for _, keysCol := range keys {
    50  			rehash(ctx, buckets, keysCol, nKeys, nil /* sel */, cancelChecker, overloadHelperVar, &datumAlloc)
    51  		}
    52  		finalizeHash(buckets, nKeys, numBuckets)
    53  	}
    54  
    55  	numKeysInSameBucket := 0
    56  	for key := range bucketsA {
    57  		if bucketsA[key] == bucketsB[key] {
    58  			numKeysInSameBucket++
    59  		}
    60  	}
    61  	// We expect that about 1/numBuckets keys remained in the same bucket, so if
    62  	// the actual number deviates by more than a factor of 3, we fail the test.
    63  	if nKeys*3/int(numBuckets) < numKeysInSameBucket {
    64  		t.Fatal(fmt.Sprintf("too many keys remained in the same bucket: expected about %d, actual %d",
    65  			nKeys/int(numBuckets), numKeysInSameBucket))
    66  	}
    67  }