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 c7fe5a974 refactor(manifest)!: remove deprecated block size builder
setter (#2325)
c7fe5a974 is described below
commit c7fe5a9747e4d7c7ee23ba80719267b5abbfb119
Author: Minh Vu <[email protected]>
AuthorDate: Tue Oct 27 07:58:34 2026 +0100
refactor(manifest)!: remove deprecated block size builder setter (#2325)
* refactor(manifest)!: remove deprecated block size builder setter
Remove the public setter for a field deprecated by the Iceberg spec.
Preserve legacy v1 manifest decoding and projection coverage.
Signed-off-by: Minh Vu <[email protected]>
* docs(release): note v1 API compatibility change
* chore: sync release notes with main to resolve conflicts
* fix(manifest): write schema default for unset v1 block sizes
* style(manifest): gofmt block size field and struct alignment
---------
Signed-off-by: Minh Vu <[email protected]>
Co-authored-by: Matt Topol <[email protected]>
---
manifest.go | 25 ++++++++++---------------
manifest_projection_test.go | 9 +++++++--
website/src/releases.md | 1 +
3 files changed, 18 insertions(+), 17 deletions(-)
diff --git a/manifest.go b/manifest.go
index 029de6eee..577e6e04a 100644
--- a/manifest.go
+++ b/manifest.go
@@ -2601,14 +2601,16 @@ func fitDecimalBytes(bytes []byte, size int) ([]byte,
error) {
}
type dataFile struct {
- Content ManifestEntryContent `avro:"content"`
- projected bool
- Path string `avro:"file_path"`
- Format FileFormat `avro:"file_format"`
- PartitionData map[string]any `avro:"partition"`
- RecordCount int64 `avro:"record_count"`
- FileSize int64 `avro:"file_size_in_bytes"`
- BlockSizeInBytes int64 `avro:"block_size_in_bytes"`
+ Content ManifestEntryContent `avro:"content"`
+ projected bool
+ Path string `avro:"file_path"`
+ Format FileFormat `avro:"file_format"`
+ PartitionData map[string]any `avro:"partition"`
+ RecordCount int64 `avro:"record_count"`
+ FileSize int64 `avro:"file_size_in_bytes"`
+ // BlockSizeInBytes is kept for reading legacy v1 manifests. When unset,
+ // v1 writes its 64 MiB schema default, while v2/v3 omit the field.
+ BlockSizeInBytes int64
`avro:"block_size_in_bytes,omitzero"`
ColSizes *[]colMap[int, int64] `avro:"column_sizes"`
ValCounts *[]colMap[int, int64] `avro:"value_counts"`
NullCounts *[]colMap[int, int64] `avro:"null_value_counts"`
@@ -3163,13 +3165,6 @@ func NewDataFileBuilder(
}, nil
}
-// BlockSizeInBytes sets the block size in bytes for the data file. Deprecated
in v2.
-func (b *DataFileBuilder) BlockSizeInBytes(size int64) *DataFileBuilder {
- b.d.BlockSizeInBytes = size
-
- return b
-}
-
// ColumnSizes sets the column sizes for the data file.
func (b *DataFileBuilder) ColumnSizes(sizes map[int]int64) *DataFileBuilder {
b.d.ColSizes = mapToAvroColMap(sizes)
diff --git a/manifest_projection_test.go b/manifest_projection_test.go
index 3195f6a61..006d9d4e6 100644
--- a/manifest_projection_test.go
+++ b/manifest_projection_test.go
@@ -20,6 +20,7 @@ package iceberg
import (
"bytes"
"reflect"
+ "strings"
"sync"
"testing"
@@ -36,6 +37,7 @@ func
TestManifestEntryProjectionWhitelistCoversDataFileSchema(t *testing.T) {
dataFileFields := make(map[string]struct{})
for field := range dataFileType.Fields() {
if name, ok := field.Tag.Lookup("avro"); ok {
+ name, _, _ = strings.Cut(name, ",")
dataFileFields[name] = struct{}{}
}
}
@@ -214,7 +216,6 @@ func
TestManifestEntryProjectionSupportsManifestVersionsAndDeletes(t *testing.T)
)
require.NoError(t, err)
builder.
- BlockSizeInBytes(64 * 1024).
ColumnSizes(map[int]int64{1: 64}).
ValueCounts(map[int]int64{1: 1}).
NullValueCounts(map[int]int64{1: 0}).
@@ -266,7 +267,7 @@ func
TestManifestEntryProjectionSupportsManifestVersionsAndDeletes(t *testing.T)
assert.Empty(t, projected.DataFile().ValueCounts())
assert.Empty(t, projected.DataFile().LowerBoundValues())
if tt.version == 1 {
- assert.Equal(t, int64(64*1024),
projected.DataFile().(*dataFile).BlockSizeInBytes)
+ assert.Equal(t, int64(64*1024*1024),
projected.DataFile().(*dataFile).BlockSizeInBytes)
}
if tt.entryContent == EntryContentPosDeletes &&
tt.format == PuffinFile {
require.NotNil(t,
projected.DataFile().ReferencedDataFile())
@@ -295,6 +296,10 @@ func
TestManifestEntryProjectionSupportsManifestVersionsAndDeletes(t *testing.T)
require.NoError(t, err)
require.NoError(t, fullReader.Close())
assert.Equal(t, map[int]int64{1: 64},
full.DataFile().ColumnSizes())
+ if tt.version == 1 {
+ // An unset builder field must serialize using
the v1 schema default.
+ assert.Equal(t, int64(64*1024*1024),
full.DataFile().(*dataFile).BlockSizeInBytes)
+ }
fs := iceio.NewMemFS()
require.NoError(t, fs.WriteFile(manifest.FilePath(),
manifestBytes.Bytes()))
diff --git a/website/src/releases.md b/website/src/releases.md
index b24209296..4cfc00695 100644
--- a/website/src/releases.md
+++ b/website/src/releases.md
@@ -42,6 +42,7 @@ Per-release notes (highlights, breaking changes,
contributors) are published on
### Unreleased compatibility notes
+- **Data file block size setter removed.**
`iceberg.DataFileBuilder.BlockSizeInBytes` is removed. New v1 manifests use the
Avro schema's 64 MiB default when the field is unset. V2/v3 omit this
deprecated field, and legacy v1 values remain readable.
- **View creation APIs take a named SQL representation.**
`view.NewVersionFromSQL`, `view.CreateView`, `view.CreateViewWithIOProperties`,
and `sql.Catalog.CreateView` now accept a `view.Representation{SQL: sql,
Dialect: dialect}` instead of positional SQL/dialect strings. The dialect is
required; Hive preserves the selected representation's dialect.
- **View SQL representation construction uses named fields.** Replace
`view.NewRepresentation(sql, dialect)` with
`view.NewRepresentation(view.Representation{SQL: sql, Dialect: dialect})`.
Serialized JSON is unchanged.
