Skip to content
This repository was archived by the owner on Nov 7, 2025. It is now read-only.

Refactor towards better table creation #1439

Merged
merged 3 commits into from
Jun 4, 2025
Merged

Refactor towards better table creation #1439

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
035d6de
Extra logging at tests
jakozaur Jun 3, 2025
0f28d7b
Remove Table.Created
jakozaur Jun 4, 2025
47ca0b9
Make gofmt happy
jakozaur Jun 4, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 1 addition & 5 deletions platform/clickhouse/clickhouse.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -318,16 +318,12 @@ func (lm *LogManager) GetTableDefinitions() (TableMap, error) {
func (lm *LogManager) AddTableIfDoesntExist(table *Table) bool {
t := lm.FindTable(table.Name)
if t == nil {
table.Created = true

table.ApplyIndexConfig(lm.cfg)

lm.tableDiscovery.TableDefinitions().Store(table.Name, table)
return true
}
wasntCreated := !t.Created
t.Created = true
return wasntCreated
return false
}

func (lm *LogManager) Ping() error {
Expand Down
6 changes: 2 additions & 4 deletions platform/clickhouse/clickhouse_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -468,8 +468,7 @@ func TestJsonConvertingBoolToStringAttr(t *testing.T) {
// Doesn't test for 100% equality, as map iteration order isn't deterministic, but should definitely be good enough.
func TestCreateTableString_1(t *testing.T) {
table := Table{
Created: false,
Name: "/_bulk?refresh=false&_source_includes=originId&require_alias=true_16",
Name: "/_bulk?refresh=false&_source_includes=originId&require_alias=true_16",
Cols: map[string]*Column{
"doc": {
Name: "doc",
Expand Down Expand Up @@ -570,8 +569,7 @@ func TestCreateTableString_1(t *testing.T) {
// Doesn't test for 100% equality, as map iteration order isn't deterministic, but should definitely be good enough.
func TestCreateTableString_NewDateTypes(t *testing.T) {
table := Table{
Created: false,
Name: "abc",
Name: "abc",
Cols: map[string]*Column{
"low_card_string": {
Name: "low_card_string",
Expand Down
1 change: 0 additions & 1 deletion platform/clickhouse/table.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,6 @@ type Table struct {
ClusterName string `default:""`
Cols map[string]*Column
Config *ChTableConfig
Created bool // do we need to create it during first insert
Indexes []IndexStatement
aliases map[string]string //deprecated
// TODO: we should use aliases directly from configuration, not store them here
Expand Down
1 change: 0 additions & 1 deletion platform/clickhouse/table_discovery.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -412,7 +412,6 @@ func (td *tableDiscovery) populateTableDefinitions(configuredTables map[string]d

if !partiallyResolved {
table := Table{
Created: true,
Name: tableName,
Comment: resTable.comment,
DatabaseName: databaseName,
Expand Down
8 changes: 4 additions & 4 deletions platform/frontend_connectors/count_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,16 +29,16 @@ func TestCountEndpoint(t *testing.T) {

tables := clickhouse.NewTableMap()
tables.Store("no_db_name", &clickhouse.Table{
Name: "no_db_name", Config: clickhouse.NewChTableConfigTimestampStringAttr(), Created: true, Cols: map[string]*clickhouse.Column{},
Name: "no_db_name", Config: clickhouse.NewChTableConfigTimestampStringAttr(), Cols: map[string]*clickhouse.Column{},
})
tables.Store("with_db_name", &clickhouse.Table{
Name: "with_db_name", Config: clickhouse.NewChTableConfigTimestampStringAttr(), Created: true, Cols: map[string]*clickhouse.Column{}, DatabaseName: "db_name",
Name: "with_db_name", Config: clickhouse.NewChTableConfigTimestampStringAttr(), Cols: map[string]*clickhouse.Column{}, DatabaseName: "db_name",
})
tables.Store("common_prefix_1", &clickhouse.Table{
Name: "common_prefix_1", Config: clickhouse.NewChTableConfigTimestampStringAttr(), Created: true, Cols: map[string]*clickhouse.Column{}, DatabaseName: "db_name",
Name: "common_prefix_1", Config: clickhouse.NewChTableConfigTimestampStringAttr(), Cols: map[string]*clickhouse.Column{}, DatabaseName: "db_name",
})
tables.Store("common_prefix_2", &clickhouse.Table{
Name: "common_prefix_2", Config: clickhouse.NewChTableConfigTimestampStringAttr(), Created: true, Cols: map[string]*clickhouse.Column{},
Name: "common_prefix_2", Config: clickhouse.NewChTableConfigTimestampStringAttr(), Cols: map[string]*clickhouse.Column{},
})

conn, mock := util.InitSqlMockWithPrettySqlAndPrint(t, false)
Expand Down
2 changes: 0 additions & 2 deletions platform/frontend_connectors/search_opensearch_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,7 +27,6 @@ func TestSearchOpensearch(t *testing.T) {
"message_____": {Name: "message_____", Type: clickhouse.NewBaseType("String")},
"__bytes": {Name: "__bytes", Type: clickhouse.NewBaseType("Int64")},
},
Created: true,
}

s := &schema.StaticRegistry{Tables: map[schema.IndexName]schema.Schema{}}
Expand Down Expand Up @@ -174,7 +173,6 @@ func TestHighlighter(t *testing.T) {
"host_name": {Name: "host_name", Type: clickhouse.NewBaseType("String")},
"_timestamp": {Name: "_timestamp", Type: clickhouse.NewBaseType("DateTime64")},
},
Created: true,
}
fields := map[schema.FieldName]schema.Field{
"messeage$*%:;": {PropertyName: "message$*%:;", InternalPropertyName: "message_____", Type: schema.QuesmaTypeText},
Expand Down
16 changes: 1 addition & 15 deletions platform/frontend_connectors/search_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -69,7 +69,6 @@ func TestAsyncSearchHandler(t *testing.T) {
Type: clickhouse.NewBaseType("String"),
},
},
Created: true,
})
fields := map[schema.FieldName]schema.Field{
"@timestamp": {PropertyName: "@timestamp", InternalPropertyName: "@timestamp", Type: schema.QuesmaTypeDate},
Expand Down Expand Up @@ -114,7 +113,6 @@ func TestAsyncSearchHandlerSpecialCharacters(t *testing.T) {
"message_____": {Name: "message_____", Type: clickhouse.NewBaseType("String")},
"__bytes": {Name: "__bytes", Type: clickhouse.NewBaseType("Int64")},
},
Created: true,
}
fields := map[schema.FieldName]schema.Field{
"host.name": {PropertyName: "host.name", InternalPropertyName: "host_name", Type: schema.QuesmaTypeObject},
Expand Down Expand Up @@ -168,7 +166,6 @@ var table = util.NewSyncMapWith(tableName, &clickhouse.Table{
Type: clickhouse.NewBaseType("String"),
},
},
Created: true,
})

func TestSearchHandler(t *testing.T) {
Expand Down Expand Up @@ -262,7 +259,6 @@ func TestSearchHandler(t *testing.T) {
Type: clickhouse.NewBaseType("String"),
},
},
Created: true,
}

tableMap := util.NewSyncMapWith(tableName, tab)
Expand Down Expand Up @@ -373,7 +369,6 @@ func TestSearchHandlerRuntimeMappings(t *testing.T) {
Type: clickhouse.NewBaseType("String"),
},
},
Created: true,
})

s := &schema.StaticRegistry{
Expand Down Expand Up @@ -425,7 +420,6 @@ func TestSearchHandlerNoAttrsConfig(t *testing.T) {
"message": {Name: "message", Type: clickhouse.NewBaseType("String")},
"@timestamp": {Name: "@timestamp", Type: clickhouse.NewBaseType("DateTime64")},
},
Created: true,
})

s := &schema.StaticRegistry{
Expand Down Expand Up @@ -477,7 +471,6 @@ func TestAsyncSearchFilter(t *testing.T) {
Type: clickhouse.BaseType{Name: "Map(String, String)"},
},
},
Created: true,
}
s := &schema.StaticRegistry{
Tables: map[schema.IndexName]schema.Schema{
Expand Down Expand Up @@ -553,7 +546,7 @@ func TestHandlingDateTimeFields(t *testing.T) {
},
}

table := clickhouse.Table{Name: tableName, Config: clickhouse.NewChTableConfigTimestampStringAttr(), Created: true,
table := clickhouse.Table{Name: tableName, Config: clickhouse.NewChTableConfigTimestampStringAttr(),
Cols: map[string]*clickhouse.Column{
"timestamp": {Name: "timestamp", Type: clickhouse.NewBaseType("DateTime")},
"timestamp64": {Name: "timestamp64", Type: clickhouse.NewBaseType("DateTime64")},
Expand Down Expand Up @@ -685,7 +678,6 @@ func TestNumericFacetsQueries(t *testing.T) {
Type: clickhouse.NewBaseType("Float64"),
},
},
Created: true,
})
handlers := []string{"handleSearch", "handleAsyncSearch"}
for i, tt := range testdata.TestsNumericFacets {
Expand Down Expand Up @@ -779,7 +771,6 @@ func TestSearchTrackTotalCount(t *testing.T) {
"message": {Name: "message", Type: clickhouse.NewBaseType("String")},
"@timestamp": {Name: "@timestamp", Type: clickhouse.NewBaseType("DateTime64")},
},
Created: true,
})

test := func(handlerName string, testcase testdata.FullSearchTestCase) {
Expand Down Expand Up @@ -885,7 +876,6 @@ func TestFullQueryTestWIP(t *testing.T) {
"message": {Name: "message", Type: clickhouse.NewBaseType("String")},
"@timestamp": {Name: "@timestamp", Type: clickhouse.NewBaseType("DateTime64")},
},
Created: true,
})

test := func(handlerName string, testcase testdata.FullSearchTestCase) {
Expand Down Expand Up @@ -996,7 +986,6 @@ func TestSearchAfterParameter_sortByJustTimestamp(t *testing.T) {
"message": {Name: "message", Type: clickhouse.NewBaseType("String")},
"@timestamp": {Name: "@timestamp", Type: clickhouse.NewBaseType("DateTime64")},
},
Created: true,
})

someTime := time.Date(2024, 1, 29, 18, 11, 36, 491000000, time.UTC) // 1706551896491 in UnixMilli
Expand Down Expand Up @@ -1133,7 +1122,6 @@ func TestSearchAfterParameter_sortByJustOneStringField(t *testing.T) {
Cols: map[string]*clickhouse.Column{
"message": {Name: "message", Type: clickhouse.NewBaseType("String")},
},
Created: true,
})

iterations := []struct {
Expand Down Expand Up @@ -1252,7 +1240,6 @@ func TestSearchAfterParameter_sortByMultipleFields(t *testing.T) {
"bicep_size": {Name: "bicep_size", Type: clickhouse.NewBaseType("Int64")},
"@timestamp": {Name: "@timestamp", Type: clickhouse.NewBaseType("DateTime64")},
},
Created: true,
})

someTime := time.Date(2024, 1, 29, 18, 11, 36, 491000000, time.UTC) // 1706551896491 in UnixMilli
Expand Down Expand Up @@ -1410,7 +1397,6 @@ func TestSearchAfterParameter_sortByNoField(t *testing.T) {
"bicep_size": {Name: "bicep_size", Type: clickhouse.NewBaseType("Int64")},
"@timestamp": {Name: "@timestamp", Type: clickhouse.NewBaseType("DateTime64")},
},
Created: true,
})

someTime := time.Date(2024, 1, 29, 18, 11, 36, 491000000, time.UTC) // 1706551896491 in UnixMilli
Expand Down
1 change: 0 additions & 1 deletion platform/frontend_connectors/terms_enum_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -94,7 +94,6 @@ func testHandleTermsEnumRequest(t *testing.T, requestBody []byte, fieldName stri
Type: clickhouse.NewBaseType("LowCardinality(String)"),
},
},
Created: true,
}
tableResolver := table_resolver.NewEmptyTableResolver()
managementConsole := ui.NewQuesmaManagementConsole(&config.QuesmaConfiguration{}, nil, make(<-chan logger.LogWithLevel, 50000), diag.EmptyPhoneHomeRecentStatsProvider(), nil, tableResolver)
Expand Down
4 changes: 1 addition & 3 deletions platform/ingest/common_table_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -171,8 +171,7 @@ func TestIngestToCommonTable(t *testing.T) {
Type: clickhouse.BaseType{Name: "Map(String, String)"},
},
},
Config: NewDefaultCHConfig(),
Created: true,
Config: NewDefaultCHConfig(),
}

for _, col := range tt.alreadyExistingColumns {
Expand Down Expand Up @@ -220,7 +219,6 @@ func TestIngestToCommonTable(t *testing.T) {
Name: indexName,
Cols: map[string]*clickhouse.Column{},
Config: NewDefaultCHConfig(),
Created: true,
VirtualTable: true,
}

Expand Down
1 change: 0 additions & 1 deletion platform/ingest/ingest_validator_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -234,7 +234,6 @@ func TestIngestValidation(t *testing.T) {
},
}},
},
Created: true,
})
for i := range inputJson {
conn, mock := util.InitSqlMockWithPrettyPrint(t, true)
Expand Down
15 changes: 5 additions & 10 deletions platform/ingest/insert_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -151,7 +151,6 @@ func ingestProcessorsNonEmpty(cfg *clickhouse.ChTableConfig) []ingestProcessorHe
"message": lowCardinalityString("message"),
"non-insert-field": genericString("non-insert-field"),
},
Created: created,
})
lms = append(lms, ingestProcessorHelper{newIngestProcessorWithEmptyTableMap(full, &config.QuesmaConfiguration{}), created})
}
Expand Down Expand Up @@ -198,7 +197,7 @@ func TestAutomaticTableCreationAtInsert(t *testing.T) {
// check if we properly create table in our tables table :) (:) suggested by Copilot) if needed
tableInMemory := ip.ip.FindTable(tableName)
needCreate := true
if tableInMemory != nil && tableInMemory.Created {
if tableInMemory != nil {
needCreate = false
}
noSuchTable := ip.ip.AddTableIfDoesntExist(table)
Expand All @@ -207,7 +206,6 @@ func TestAutomaticTableCreationAtInsert(t *testing.T) {
// and Created is set to true
tableInMemory = ip.ip.FindTable(tableName)
assert.NotNil(t, tableInMemory)
assert.True(t, tableInMemory.Created)

// and we have a schema in memory in every case
assert.Equal(t, 1, ip.ip.tableDiscovery.TableDefinitions().Size())
Expand Down Expand Up @@ -253,9 +251,7 @@ func TestProcessInsertQuery(t *testing.T) {
// info: result values aren't important, this '.WillReturnResult[...]' just needs to be there
if !ip.tableAlreadyCreated {
// we check here if we try to create table from predefined schema, not from insert's JSON
if ip.ip.tableDiscovery.TableDefinitions().Size() > 0 {
mock.ExpectExec(`CREATE TABLE IF NOT EXISTS "` + tableName + `.*non-insert-field`).WillReturnResult(sqlmock.NewResult(0, 0))
} else {
if ip.ip.tableDiscovery.TableDefinitions().Size() == 0 {
mock.ExpectExec(`CREATE TABLE IF NOT EXISTS "` + tableName).WillReturnResult(sqlmock.NewResult(0, 0))
}
}
Expand Down Expand Up @@ -311,10 +307,9 @@ func TestInsertVeryBigIntegers(t *testing.T) {

// big integer as an attribute field
tableMapNoSchemaFields := util.NewSyncMapWith(tableName, &clickhouse.Table{
Name: tableName,
Config: NewChTableConfigFourAttrs(),
Cols: map[string]*clickhouse.Column{},
Created: true,
Name: tableName,
Config: NewChTableConfigFourAttrs(),
Cols: map[string]*clickhouse.Column{},
})

for i, bigInt := range bigInts {
Expand Down
21 changes: 21 additions & 0 deletions platform/ingest/parserCreateTable_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -364,3 +364,24 @@ func TestParseCreateTableWithNullable(t *testing.T) {
assert.True(t, table.Cols["array-tuple"].Type.(clickhouse.CompoundType).BaseType.(clickhouse.MultiValueType).Cols[0].Type.IsNullable())
assert.False(t, table.Cols["array-tuple"].Type.(clickhouse.CompoundType).BaseType.(clickhouse.MultiValueType).Cols[1].Type.IsNullable())
}

func IgnoredTestParseCreateTableWithDots(t *testing.T) {
const columnNr = 5
q := `CREATE TABLE IF NOT EXISTS "my-index-2.3"
(
"attributes_values" Map(String,String),
"attributes_metadata" Map(String,String),

"level" Nullable(String) COMMENT 'quesmaMetadataV1:fieldName=level',
"@timestamp" DateTime64 DEFAULT now64() COMMENT 'quesmaMetadataV1:fieldName=%40timestamp',
"message" Nullable(String) COMMENT 'quesmaMetadataV1:fieldName=message'
)
ENGINE = MergeTree
ORDER BY ("@timestamp")

COMMENT 'created by Quesma'`
table, err := clickhouse.NewTable(q, nil)
assert.NoError(t, err)
assert.Equal(t, columnNr, len(table.Cols))
assert.Equal(t, "my-index-2.3", table.Name)
}
19 changes: 9 additions & 10 deletions platform/ingest/processor.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -667,11 +667,11 @@ func (ip *IngestProcessor) processInsertQuery(ctx context.Context,
if err != nil {
logger.ErrorWithCtx(ctx).Msgf("error createTableObjectAndAttributes, can't create table: %v", err)
return nil, err
} else {
logger.InfoWithCtx(ctx).Msgf("created table '%s' with query: %s", tableName, createTableCmd)
}
// Set pointer to table after creating it
table = ip.FindTable(tableName)
} else if !table.Created {
createTableCmd = table.CreateTableString()
}
if table == nil {
return nil, fmt.Errorf("table %s not found", tableName)
Expand Down Expand Up @@ -824,15 +824,18 @@ func (ip *IngestProcessor) processInsertQueryInternal(ctx context.Context, table
tableFormatter TableColumNameFormatter, isVirtualTable bool) error {
statements, err := ip.processInsertQuery(ctx, tableName, jsonData, transformer, tableFormatter, isVirtualTable)
if err != nil {
logger.ErrorWithCtx(ctx).Msgf("error processing insert query: %v", err)
return err
}

var logVirtualTableDDL bool // maybe this should be a part of the config or sth

if isVirtualTable && logVirtualTableDDL {
for _, statement := range statements {
if strings.HasPrefix(statement, "ALTER") || strings.HasPrefix(statement, "CREATE") {
for _, statement := range statements {
if strings.HasPrefix(statement, "ALTER") || strings.HasPrefix(statement, "CREATE") {
if isVirtualTable && logVirtualTableDDL {
logger.InfoWithCtx(ctx).Msgf("VIRTUAL DDL EXECUTION: %s", statement)
} else {
logger.InfoWithCtx(ctx).Msgf("DDL EXECUTION: %s", statement)
}
}
}
Expand Down Expand Up @@ -975,8 +978,6 @@ func (ip *IngestProcessor) storeVirtualTable(table *chLib.Table) error {
func (ip *IngestProcessor) AddTableIfDoesntExist(table *chLib.Table) bool {
t := ip.FindTable(table.Name)
if t == nil {
table.Created = true

table.ApplyIndexConfig(ip.cfg)

if table.VirtualTable {
Expand All @@ -988,9 +989,7 @@ func (ip *IngestProcessor) AddTableIfDoesntExist(table *chLib.Table) bool {
ip.tableDiscovery.AddTable(table.Name, table)
return true
}
wasntCreated := !t.Created
t.Created = true
return wasntCreated
return false
}

func (ip *IngestProcessor) GetSchemaRegistry() schema.Registry {
Expand Down
6 changes: 2 additions & 4 deletions platform/ingest/processor_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -604,8 +604,7 @@ func TestJsonConvertingBoolToStringAttr(t *testing.T) {
// Doesn't test for 100% equality, as map iteration order isn't deterministic, but should definitely be good enough.
func TestCreateTableString_1(t *testing.T) {
table := clickhouse.Table{
Created: false,
Name: "/_bulk?refresh=false&_source_includes=originId&require_alias=true_16",
Name: "/_bulk?refresh=false&_source_includes=originId&require_alias=true_16",
Cols: map[string]*clickhouse.Column{
"doc": {
Name: "doc",
Expand Down Expand Up @@ -706,8 +705,7 @@ func TestCreateTableString_1(t *testing.T) {
// Doesn't test for 100% equality, as map iteration order isn't deterministic, but should definitely be good enough.
func TestCreateTableString_NewDateTypes(t *testing.T) {
table := clickhouse.Table{
Created: false,
Name: "abc",
Name: "abc",
Cols: map[string]*clickhouse.Column{
"low_card_string": {
Name: "low_card_string",
Expand Down
Loading
Loading