utils

func AnalyzeTable ¶

func AnalyzeTable(ctx context.Context, db *sql.DB, tableName string) error

AnalyzeTable do 'ANALYZE TABLE `table`' SQL.

func CalculateChunkSize ¶

func CalculateChunkSize(rowCount int64) int64

CalculateChunkSize returns chunkSize according to table rows count.

func CompareData ¶

func CompareData(map1, map2 map[string]*dbutil.ColumnData, orderKeyCols, columns []*model.ColumnInfo) (equal bool, cmp int32, err error)

CompareData compare two row datas. equal = true: map1 = map2 equal = false:

1. cmp = 0: map1 and map2 have the same orderkeycolumns, but other columns are in difference.
2. cmp = -1: map1 < map2 (by comparing the orderkeycolumns)
3. cmp = 1: map1 > map2

func CompareStruct ¶

func CompareStruct(upstreamTableInfos []*model.TableInfo, downstreamTableInfo *model.TableInfo) (isEqual bool, isPanic bool)

CompareStruct compare tables' columns and indices from upstream and downstream. There are 2 return values:

isEqual	: result of comparing tables' columns and indices
isPanic	: the differences of tables' struct can not be ignored. Need to skip data comparing.

func GenerateDeleteDML ¶

func GenerateDeleteDML(data map[string]*dbutil.ColumnData, table *model.TableInfo, schema string) string

GerateReplaceDMLWithAnnotation returns the delete SQL for the specific row.

func GenerateReplaceDML ¶

func GenerateReplaceDML(data map[string]*dbutil.ColumnData, table *model.TableInfo, schema string) string

GenerateReplaceDML returns the insert SQL for the specific row values.

func GenerateReplaceDMLWithAnnotation ¶

func GenerateReplaceDMLWithAnnotation(source, target map[string]*dbutil.ColumnData, table *model.TableInfo, schema string) string

GerateReplaceDMLWithAnnotation returns the replace SQL for the specific 2 rows. And add Annotations to show the different columns.

func GetApproximateMidBySize ¶

func GetApproximateMidBySize(ctx context.Context, db *sql.DB, schema, table string, indexColumns []*model.ColumnInfo, limitRange string, args []interface{}, count int64) (map[string]string, error)

GetApproximateMidBySize return the `count`th row in rows that meet the `limitRange`.

func GetBetterIndex ¶

func GetBetterIndex(ctx context.Context, db *sql.DB, schema, table string, tableInfo *model.TableInfo) ([]*model.IndexInfo, error)

GetBetterIndex returns the index more dinstict. If the index is primary key or unique, it can be return directly. Otherwise select the index which has higher value of `COUNT(DISTINCT a)/COUNT(*)`.

func GetChunkIDFromSQLFileName ¶

func GetChunkIDFromSQLFileName(fileIDStr string) (int, int, int, int, error)

GetChunkIDFromSQLFileName convert the filename to chunk's `Index`.

func GetColumnsFromIndex ¶

func GetColumnsFromIndex(index *model.IndexInfo, tableInfo *model.TableInfo) []*model.ColumnInfo

GetColumnsFromIndex returns `ColumnInfo`s of the specified index.

func GetCountAndCRC32Checksum ¶

func GetCountAndCRC32Checksum(ctx context.Context, db *sql.DB, schemaName, tableName string, tbInfo *model.TableInfo, limitRange string, args []interface{}) (int64, int64, error)

GetCountAndCRC32Checksum returns checksum code and count of some data by given condition

func GetPDAddrs ¶

func GetPDAddrs(ctx context.Context, db *sql.DB) ([]string, error)

GetPDAddrs gets PD address from TiDB

func GetPDClientForGC ¶

func GetPDClientForGC(ctx context.Context, db *sql.DB) (pd.Client, error)

GetPDClientForGC is an initialization step.

func GetSQLFileName ¶

func GetSQLFileName(index *chunk.ChunkID) string

GetSQLFileName returns filename of fix-SQL identified by chunk's `Index`.

func GetSelectivity ¶

func GetSelectivity(ctx context.Context, db *sql.DB, schemaName, tableName, columnName string, tbInfo *model.TableInfo) (float64, error)

GetSelectivity returns the value of `COUNT(DISTINCT col)/COUNT(1)` SQL.

func GetSnapshot ¶

func GetSnapshot(ctx context.Context, db *sql.DB) ([]string, error)

func GetSpecifiedColumnValueAndClose ¶

func GetSpecifiedColumnValueAndClose(rows *sql.Rows, columnName string) ([]string, error)

GetSpecifiedColumnValueAndClose get columns' values whose name is equal to columnName and close the given rows

func GetTableRowsQueryFormat ¶

func GetTableRowsQueryFormat(schema, table string, tableInfo *model.TableInfo, collation string) (string, []*model.ColumnInfo)

GetTableRowsQueryFormat returns a rowsQuerySQL template for the specific table.

e.g. SELECT /*!40001 SQL_NO_CACHE */ `a`, `b` FROM `schema`.`table` WHERE %s ORDER BY `a`.

func GetTableSize ¶

func GetTableSize(ctx context.Context, db *sql.DB, schemaName, tableName string) (int64, error)

GetTableSize loads the TableSize from `information_schema`.`tables`.

func MinLenInSlices ¶

func MinLenInSlices(slices [][]string) int

MinLenInSlices returns the smallest length among slices.

func NeedQuotes ¶

func NeedQuotes(tp byte) bool

NeedQuotes determines whether an escape character is required for `'`.

func ResetColumns ¶

func ResetColumns(tableInfo *model.TableInfo, columns []string) (*model.TableInfo, bool)

ResetColumns removes index from `tableInfo.Indices`, whose columns appear in `columns`. And removes column from `tableInfo.Columns`, which appears in `columns`. And initializes the offset of the column of each index to new `tableInfo.Columns`.

Return the new tableInfo and the flag whether the columns have timestamp type.

func SliceToMap ¶

func SliceToMap(slice []string) map[string]interface{}

SliceToMap converts Slice to Set

func StartGCSavepointUpdateService ¶

func StartGCSavepointUpdateService(ctx context.Context, pdCli pd.Client, db *sql.DB, snapshot string) error

StartGCSavepointUpdateService keeps GC safePoint stop moving forward.

func UniqueID ¶

func UniqueID(schema string, table string) string

UniqueID returns `schema:table`