This is an automated email from the ASF dual-hosted git repository.
zeroshade pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/iceberg-go.git
The following commit(s) were added to refs/heads/main by this push:
new a4fb4d0da refactor: export the position delete file_path and pos field
IDs (#2245)
a4fb4d0da is described below
commit a4fb4d0dafaeaf455990c1c3142462107662e5ec
Author: Revanth <[email protected]>
AuthorDate: Wed Oct 21 17:45:46 2026 -0500
refactor: export the position delete file_path and pos field IDs (#2245)
---
manifest.go | 4 ++--
metadata_columns.go | 11 +++++++++++
table/data_file_meta.go | 6 ++----
table/inspect_position_deletes.go | 4 ++--
table/internal/parquet_files.go | 6 +-----
5 files changed, 18 insertions(+), 13 deletions(-)
diff --git a/manifest.go b/manifest.go
index 08cb21011..eed525b1d 100644
--- a/manifest.go
+++ b/manifest.go
@@ -3361,6 +3361,6 @@ type ManifestEntry interface {
}
var PositionalDeleteSchema = NewSchema(0,
- NestedField{ID: 2147483546, Type: PrimitiveTypes.String, Name:
"file_path", Required: true},
- NestedField{ID: 2147483545, Type: PrimitiveTypes.Int64, Name: "pos",
Required: true},
+ NestedField{ID: DeleteFilePathFieldID, Type: PrimitiveTypes.String,
Name: "file_path", Required: true},
+ NestedField{ID: DeleteFilePosFieldID, Type: PrimitiveTypes.Int64, Name:
"pos", Required: true},
)
diff --git a/metadata_columns.go b/metadata_columns.go
index a1b509a55..3e2fe4577 100644
--- a/metadata_columns.go
+++ b/metadata_columns.go
@@ -43,6 +43,17 @@ const (
LastUpdatedSequenceNumberFieldID = 2147483539
)
+// Position delete file column field IDs. Reserved IDs are Integer.MAX_VALUE -
101
+// and Integer.MAX_VALUE - 102 per the Iceberg spec (Position Delete Files).
+const (
+ // DeleteFilePathFieldID is the field ID for file_path (required
string), the
+ // data file a position delete applies to. Reserved as
Integer.MAX_VALUE - 101.
+ DeleteFilePathFieldID = 2147483546
+ // DeleteFilePosFieldID is the field ID for pos (required long), the
ordinal
+ // of the deleted row in that file. Reserved as Integer.MAX_VALUE - 102.
+ DeleteFilePosFieldID = 2147483545
+)
+
// Row lineage metadata column names (v3+).
const (
RowIDColumnName = "_row_id"
diff --git a/table/data_file_meta.go b/table/data_file_meta.go
index 7b8e116ed..99a5fdff4 100644
--- a/table/data_file_meta.go
+++ b/table/data_file_meta.go
@@ -28,8 +28,6 @@ import (
tblutils "github.com/apache/iceberg-go/table/internal"
)
-const referencedDataFilePathFieldID = 2147483546
-
// DataFileArgs collects everything needed to compute a fully populated
// [iceberg.DataFile] (data, equality-delete, or positional-delete) from
// a file's in-memory, format-specific metadata object without performing
@@ -255,7 +253,7 @@ func DataFileFromMetadata(args DataFileArgs)
(iceberg.DataFile, error) {
if statsProps == nil {
statsProps = iceberg.Properties{}
}
- if colName, ok :=
args.Schema.FindColumnName(referencedDataFilePathFieldID); ok {
+ if colName, ok :=
args.Schema.FindColumnName(iceberg.DeleteFilePathFieldID); ok {
statsProps[MetricsModeColumnConfPrefix+"."+colName] =
string(tblutils.MetricModeFull)
}
}
@@ -335,7 +333,7 @@ func DataFileFromMetadata(args DataFileArgs)
(iceberg.DataFile, error) {
func verifyReferencedDataFile(df iceberg.DataFile, ref string) error {
decode := func(bounds map[int][]byte, which string) (string, error) {
- raw, ok := bounds[referencedDataFilePathFieldID]
+ raw, ok := bounds[iceberg.DeleteFilePathFieldID]
if !ok {
return "", fmt.Errorf(
"referenced_data_file %q is set but the
position-delete file_path column has no %s bound",
diff --git a/table/inspect_position_deletes.go
b/table/inspect_position_deletes.go
index 924aef562..a4f55ba4e 100644
--- a/table/inspect_position_deletes.go
+++ b/table/inspect_position_deletes.go
@@ -35,8 +35,8 @@ import (
)
const (
- positionDeleteFilePathID = math.MaxInt32 - 101
- positionDeletePosID = math.MaxInt32 - 102
+ positionDeleteFilePathID = iceberg.DeleteFilePathFieldID
+ positionDeletePosID = iceberg.DeleteFilePosFieldID
positionDeleteRowID = math.MaxInt32 - 103
positionDeletePartitionID = math.MaxInt32 - 5
positionDeleteSpecID = math.MaxInt32 - 4
diff --git a/table/internal/parquet_files.go b/table/internal/parquet_files.go
index e023739ce..eb714a07d 100644
--- a/table/internal/parquet_files.go
+++ b/table/internal/parquet_files.go
@@ -1280,10 +1280,6 @@ func (w *ParquetFileWriter) Close() (_ iceberg.DataFile,
err error) {
return stats.ToDataFile(opts), nil
}
-// posDeleteFilePathFieldID is the reserved field ID of the file_path column
-// in a position delete file.
-const posDeleteFilePathFieldID = 2147483546
-
// singleReferencedDataFile returns the data file path when every delete in a
// position delete file targets that one file, or nil otherwise. It compares
// the untruncated file_path min and max, which position delete writers always
@@ -1291,7 +1287,7 @@ const posDeleteFilePathFieldID = 2147483546
// PositionDeleteWriter leaves the field unset; the spec allows it on position
// delete files with a single target.
func singleReferencedDataFile(stats *DataFileStatistics) *string {
- agg, ok := stats.ColAggs[posDeleteFilePathFieldID]
+ agg, ok := stats.ColAggs[iceberg.DeleteFilePathFieldID]
if !ok {
return nil
}
