kylesoda/go-migrate
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

refactor: replace specific extractor implementations with a generic extractor; remove mssql and postgres extractor files

Browse Source
This commit is contained in:
Kylesoda
2026-04-26 19:33:59 -05:00
parent 33af391986
commit 9a00d6af04
5 changed files with 138 additions and 372 deletions
Download Patch File Download Diff File Expand all files Collapse all files

166
internal/app/etl/extractors/main.go
View File

@@ -12,13 +12,144 @@ import (
"git.ksdemosapps.com/kylesoda/go-migrate/internal/app/config"
"git.ksdemosapps.com/kylesoda/go-migrate/internal/app/convert"
"git.ksdemosapps.com/kylesoda/go-migrate/internal/app/custom_errors"
"git.ksdemosapps.com/kylesoda/go-migrate/internal/app/etl"
dbwrapper "git.ksdemosapps.com/kylesoda/go-migrate/internal/app/db-wrapper"
"git.ksdemosapps.com/kylesoda/go-migrate/internal/app/models"
"github.com/google/uuid"
)
func Consume(
type GenericExtractor struct {
db dbwrapper.DbWrapper
}
func NewExtractor(db dbwrapper.DbWrapper) GenericExtractor {
return GenericExtractor{db: db}
}
func errorFromLastRow(
lastRow models.UnknownRowValues,
indexPrimaryKey int,
partition models.Partition,
previousError error,
) error {
lastIdRawValue := lastRow[indexPrimaryKey]
lastId, ok := convert.ToInt64(lastIdRawValue)
if !ok {
currentPartition := partition
currentPartition.RetryCounter = 3
return &custom_errors.ExtractorError{
Partition: currentPartition,
HasLastId: true,
Msg: fmt.Sprintf("Couldn't cast last id value as int: %s", previousError.Error()),
}
}
return &custom_errors.ExtractorError{
Partition: partition,
HasLastId: true,
LastId: lastId,
Msg: previousError.Error(),
}
}
func (ex *GenericExtractor) ProcessPartition(
ctx context.Context,
tableInfo config.SourceTableInfo,
columns []models.ColumnType,
batchSize int,
partition models.Partition,
indexPrimaryKey int,
chBatchesOut chan<- models.Batch,
) (int, error) {
rowsRead := 0
query := dbwrapper.ExtractionQuery{
Schema: tableInfo.Schema,
Table: tableInfo.Table,
PrimaryKey: tableInfo.PrimaryKey,
LowerLimit: dbwrapper.ExtractorQueryLimit{
IsValid: partition.HasRange && partition.Range.Min > 0,
IsInclusive: partition.Range.IsMinInclusive,
Value: partition.Range.Min,
},
UpperLimit: dbwrapper.ExtractorQueryLimit{
IsValid: partition.HasRange && partition.Range.Max > 0,
IsInclusive: partition.Range.IsMaxInclusive,
Value: partition.Range.Max,
},
}
rows, err := ex.db.QueryFromObject(ctx, query)
if err != nil {
return rowsRead, &custom_errors.ExtractorError{Partition: partition, HasLastId: false, Msg: err.Error()}
}
defer rows.Close()
batchRows := make([]models.UnknownRowValues, 0, batchSize)
for rows.Next() {
rowValues := make([]any, len(columns))
scanArgs := make([]any, len(columns))
for i := range rowValues {
scanArgs[i] = &rowValues[i]
}
if err := rows.Scan(scanArgs...); err != nil {
if len(batchRows) == 0 {
return rowsRead, &custom_errors.ExtractorError{Partition: partition, HasLastId: false, Msg: err.Error()}
}
lastRow := batchRows[len(batchRows)-1]
select {
case chBatchesOut <- models.Batch{Id: uuid.New(), PartitionId: partition.Id, Rows: batchRows, RetryCounter: 0}:
case <-ctx.Done():
return rowsRead, ctx.Err()
}
return rowsRead, errorFromLastRow(lastRow, indexPrimaryKey, partition, err)
}
rowsRead++
batchRows = append(batchRows, rowValues)
if len(batchRows) >= batchSize {
select {
case chBatchesOut <- models.Batch{Id: uuid.New(), PartitionId: partition.Id, Rows: batchRows, RetryCounter: 0}:
case <-ctx.Done():
return rowsRead, ctx.Err()
}
batchRows = make([]models.UnknownRowValues, 0, batchSize)
}
}
if err := rows.Err(); err != nil {
if errors.Is(err, ctx.Err()) {
return rowsRead, ctx.Err()
}
if len(batchRows) > 0 {
lastRow := batchRows[len(batchRows)-1]
return rowsRead, errorFromLastRow(lastRow, indexPrimaryKey, partition, err)
}
return rowsRead, &custom_errors.ExtractorError{Partition: partition, HasLastId: false, Msg: err.Error()}
}
if len(batchRows) > 0 {
select {
case chBatchesOut <- models.Batch{Id: uuid.New(), PartitionId: partition.Id, Rows: batchRows, RetryCounter: 0}:
case <-ctx.Done():
return rowsRead, ctx.Err()
}
}
return rowsRead, nil
}
func (ex *GenericExtractor) Consume(
ctx context.Context,
extractor etl.Extractor,
tableInfo config.SourceTableInfo,
columns []models.ColumnType,
batchSize int,
@@ -59,7 +190,7 @@ func Consume(
return
}
rowsReadResult, err := extractor.ProcessPartition(
rowsReadResult, err := ex.ProcessPartition(
ctx,
tableInfo,
columns,
@@ -101,30 +232,3 @@ func Consume(
}
}
}
func errorFromLastRow(
lastRow models.UnknownRowValues,
indexPrimaryKey int,
partition models.Partition,
previousError error,
) error {
lastIdRawValue := lastRow[indexPrimaryKey]
lastId, ok := convert.ToInt64(lastIdRawValue)
if !ok {
currentPartition := partition
currentPartition.RetryCounter = 3
return &custom_errors.ExtractorError{
Partition: currentPartition,
HasLastId: true,
Msg: fmt.Sprintf("Couldn't cast last id value as int: %s", previousError.Error()),
}
}
return &custom_errors.ExtractorError{
Partition: partition,
HasLastId: true,
LastId: lastId,
Msg: previousError.Error(),
}
}

180
internal/app/etl/extractors/mssql.go
View File

@@ -1,180 +0,0 @@
package extractors
import (
"context"
"database/sql"
"errors"
"fmt"
"strings"
"git.ksdemosapps.com/kylesoda/go-migrate/internal/app/config"
"git.ksdemosapps.com/kylesoda/go-migrate/internal/app/custom_errors"
dbwrapper "git.ksdemosapps.com/kylesoda/go-migrate/internal/app/db-wrapper"
"git.ksdemosapps.com/kylesoda/go-migrate/internal/app/etl"
"git.ksdemosapps.com/kylesoda/go-migrate/internal/app/models"
"github.com/google/uuid"
)
type MssqlExtractor struct {
db dbwrapper.DbWrapper
}
func NewMssqlExtractor(db dbwrapper.DbWrapper) etl.Extractor {
return &MssqlExtractor{db: db}
}
func buildExtractQueryMssql(
tableInfo config.SourceTableInfo,
columns []models.ColumnType,
includeRange bool,
isMinInclusive bool,
isMaxInclusive bool,
hasMin bool,
hasMax bool,
) string {
var sbQuery strings.Builder
sbQuery.WriteString("SELECT ")
if len(columns) == 0 {
sbQuery.WriteString("*")
} else {
for i, col := range columns {
fmt.Fprintf(&sbQuery, "[%s]", col.Name())
if col.Type() == "GEOMETRY" {
fmt.Fprintf(&sbQuery, ".STAsBinary() AS [%s]", col.Name())
}
if i < len(columns)-1 {
sbQuery.WriteString(", ")
}
}
}
fmt.Fprintf(&sbQuery, " FROM [%s].[%s]", tableInfo.Schema, tableInfo.Table)
if includeRange && (hasMin || hasMax) {
sbQuery.WriteString(" WHERE ")
if hasMin {
fmt.Fprintf(&sbQuery, "[%s]", tableInfo.PrimaryKey)
if isMinInclusive {
sbQuery.WriteString(" >=")
} else {
sbQuery.WriteString(" >")
}
sbQuery.WriteString(" @min")
}
if hasMin && hasMax {
sbQuery.WriteString(" AND ")
}
if hasMax {
fmt.Fprintf(&sbQuery, "[%s]", tableInfo.PrimaryKey)
if isMaxInclusive {
sbQuery.WriteString(" <=")
} else {
sbQuery.WriteString(" <")
}
sbQuery.WriteString(" @max")
}
}
fmt.Fprintf(&sbQuery, " ORDER BY [%s] ASC", tableInfo.PrimaryKey)
return sbQuery.String()
}
func (mssqlEx *MssqlExtractor) ProcessPartition(
ctx context.Context,
tableInfo config.SourceTableInfo,
columns []models.ColumnType,
batchSize int,
partition models.Partition,
indexPrimaryKey int,
chBatchesOut chan<- models.Batch,
) (int, error) {
hasMin := partition.HasRange && partition.Range.Min > 0
hasMax := partition.HasRange && partition.Range.Max > 0
query := buildExtractQueryMssql(tableInfo, columns, partition.HasRange, partition.Range.IsMinInclusive, partition.Range.IsMaxInclusive, hasMin, hasMax)
var queryArgs []any
if hasMin {
queryArgs = append(queryArgs, sql.Named("min", partition.Range.Min))
}
if hasMax {
queryArgs = append(queryArgs, sql.Named("max", partition.Range.Max))
}
rowsRead := 0
rows, err := mssqlEx.db.Query(ctx, query, queryArgs...)
if err != nil {
return rowsRead, &custom_errors.ExtractorError{Partition: partition, HasLastId: false, Msg: err.Error()}
}
defer rows.Close()
batchRows := make([]models.UnknownRowValues, 0, batchSize)
for rows.Next() {
rowValues := make([]any, len(columns))
scanArgs := make([]any, len(columns))
for i := range rowValues {
scanArgs[i] = &rowValues[i]
}
if err := rows.Scan(scanArgs...); err != nil {
if len(batchRows) == 0 {
return rowsRead, &custom_errors.ExtractorError{Partition: partition, HasLastId: false, Msg: err.Error()}
}
lastRow := batchRows[len(batchRows)-1]
select {
case chBatchesOut <- models.Batch{Id: uuid.New(), PartitionId: partition.Id, Rows: batchRows, RetryCounter: 0}:
case <-ctx.Done():
return rowsRead, ctx.Err()
}
return rowsRead, errorFromLastRow(lastRow, indexPrimaryKey, partition, err)
}
rowsRead++
batchRows = append(batchRows, rowValues)
if len(batchRows) >= batchSize {
select {
case chBatchesOut <- models.Batch{Id: uuid.New(), PartitionId: partition.Id, Rows: batchRows, RetryCounter: 0}:
case <-ctx.Done():
return rowsRead, ctx.Err()
}
batchRows = make([]models.UnknownRowValues, 0, batchSize)
}
}
if err := rows.Err(); err != nil {
if errors.Is(err, ctx.Err()) {
return rowsRead, ctx.Err()
}
if len(batchRows) > 0 {
lastRow := batchRows[len(batchRows)-1]
return rowsRead, errorFromLastRow(lastRow, indexPrimaryKey, partition, err)
}
return rowsRead, &custom_errors.ExtractorError{Partition: partition, HasLastId: false, Msg: err.Error()}
}
if len(batchRows) > 0 {
select {
case chBatchesOut <- models.Batch{Id: uuid.New(), PartitionId: partition.Id, Rows: batchRows, RetryCounter: 0}:
case <-ctx.Done():
return rowsRead, ctx.Err()
}
}
return rowsRead, nil
}

157
internal/app/etl/extractors/postgres.go
View File

@@ -1,157 +0,0 @@
package extractors
import (
"context"
"fmt"
"strings"
"git.ksdemosapps.com/kylesoda/go-migrate/internal/app/config"
"git.ksdemosapps.com/kylesoda/go-migrate/internal/app/custom_errors"
dbwrapper "git.ksdemosapps.com/kylesoda/go-migrate/internal/app/db-wrapper"
"git.ksdemosapps.com/kylesoda/go-migrate/internal/app/etl"
"git.ksdemosapps.com/kylesoda/go-migrate/internal/app/models"
"github.com/google/uuid"
)
type PostgresExtractor struct {
db dbwrapper.DbWrapper
}
func NewPostgresExtractor(db dbwrapper.DbWrapper) etl.Extractor {
return &PostgresExtractor{db: db}
}
func buildExtractQueryPostgres(
sourceDbInfo config.SourceTableInfo,
columns []models.ColumnType,
includeRange bool,
isMinInclusive bool,
isMaxInclusive bool,
hasMin bool,
hasMax bool,
) string {
var sbColumns strings.Builder
if len(columns) == 0 {
sbColumns.WriteString("*")
} else {
for i, col := range columns {
if col.Type() == "GEOMETRY" {
sbColumns.WriteString(`ST_AsEWKB("`)
sbColumns.WriteString(col.Name())
sbColumns.WriteString(`") AS "`)
sbColumns.WriteString(col.Name())
sbColumns.WriteString(`"`)
} else {
sbColumns.WriteString(`"`)
sbColumns.WriteString(col.Name())
sbColumns.WriteString(`"`)
}
if i < len(columns)-1 {
sbColumns.WriteString(", ")
}
}
}
query := fmt.Sprintf(`SELECT %s FROM "%s"."%s"`, sbColumns.String(), sourceDbInfo.Schema, sourceDbInfo.Table)
if includeRange && (hasMin || hasMax) {
query += " WHERE "
paramIdx := 1
if hasMin {
query += fmt.Sprintf(`"%s"`, sourceDbInfo.PrimaryKey)
if isMinInclusive {
query += " >="
} else {
query += " >"
}
query += fmt.Sprintf(" $%d", paramIdx)
paramIdx++
}
if hasMin && hasMax {
query += " AND "
}
if hasMax {
query += fmt.Sprintf(`"%s"`, sourceDbInfo.PrimaryKey)
if isMaxInclusive {
query += " <="
} else {
query += " <"
}
query += fmt.Sprintf(" $%d", paramIdx)
}
}
query += fmt.Sprintf(` ORDER BY "%s" ASC`, sourceDbInfo.PrimaryKey)
return query
}
func (postgresEx *PostgresExtractor) ProcessPartition(
ctx context.Context,
tableInfo config.SourceTableInfo,
columns []models.ColumnType,
batchSize int,
partition models.Partition,
indexPrimaryKey int,
chBatchesOut chan<- models.Batch,
) (int, error) {
hasMin := partition.HasRange && partition.Range.Min > 0
hasMax := partition.HasRange && partition.Range.Max > 0
query := buildExtractQueryPostgres(tableInfo, columns, partition.HasRange, partition.Range.IsMinInclusive, partition.Range.IsMaxInclusive, hasMin, hasMax)
var queryArgs []any
if hasMin {
queryArgs = append(queryArgs, partition.Range.Min)
}
if hasMax {
queryArgs = append(queryArgs, partition.Range.Max)
}
rowsRead := 0
rows, err := postgresEx.db.Query(ctx, query, queryArgs...)
if err != nil {
return rowsRead, &custom_errors.ExtractorError{Partition: partition, HasLastId: false, Msg: err.Error()}
}
defer rows.Close()
batchRows := make([]models.UnknownRowValues, 0, batchSize)
for rows.Next() {
values, err := rows.Values()
if err != nil {
return rowsRead, &custom_errors.ExtractorError{Partition: partition, HasLastId: false, Msg: err.Error()}
}
rowsRead++
batchRows = append(batchRows, values)
if len(batchRows) >= batchSize {
select {
case chBatchesOut <- models.Batch{Id: uuid.New(), PartitionId: partition.Id, Rows: batchRows, RetryCounter: 0}:
case <-ctx.Done():
return rowsRead, ctx.Err()
}
batchRows = make([]models.UnknownRowValues, 0, batchSize)
}
}
if err := rows.Err(); err != nil {
return rowsRead, &custom_errors.ExtractorError{Partition: partition, HasLastId: false, Msg: err.Error()}
}
if len(batchRows) > 0 {
select {
case chBatchesOut <- models.Batch{Id: uuid.New(), PartitionId: partition.Id, Rows: batchRows, RetryCounter: 0}:
case <-ctx.Done():
return rowsRead, nil
}
}
return rowsRead, nil
}