From 6c8431900ec978cea3ef19977ca240ff92be549b Mon Sep 17 00:00:00 2001
From: Victor Perez <vperez@timescale.com>
Date: Wed, 7 May 2025 17:36:11 +0200
Subject: [PATCH] add TestDeltaByteArray test and update parquet-testing

This is intended to compare with v18 in arrow-go repository to prove that delta_byte_array.parquet file cannot be read successfully
---
 cpp/submodules/parquet-testing      |  2 +-
 go/parquet/file/file_reader_test.go | 53 +++++++++++++++++++++++++++++
 2 files changed, 54 insertions(+), 1 deletion(-)

diff --git a/cpp/submodules/parquet-testing b/cpp/submodules/parquet-testing
index 74278bc4a11..39b91cf8530 160000
--- a/cpp/submodules/parquet-testing
+++ b/cpp/submodules/parquet-testing
@@ -1 +1 @@
-Subproject commit 74278bc4a1122d74945969e6dec405abd1533ec3
+Subproject commit 39b91cf853062d92f0d20581d37b20dabe70a6a0
diff --git a/go/parquet/file/file_reader_test.go b/go/parquet/file/file_reader_test.go
index 8056a837ea1..cd381beeb22 100644
--- a/go/parquet/file/file_reader_test.go
+++ b/go/parquet/file/file_reader_test.go
@@ -18,8 +18,10 @@ package file_test
 
 import (
 	"bytes"
+	"context"
 	"crypto/rand"
 	"encoding/binary"
+	"encoding/csv"
 	"io"
 	"os"
 	"path"
@@ -34,6 +36,7 @@ import (
 	format "github.com/apache/arrow/go/v17/parquet/internal/gen-go/parquet"
 	"github.com/apache/arrow/go/v17/parquet/internal/thrift"
 	"github.com/apache/arrow/go/v17/parquet/metadata"
+	"github.com/apache/arrow/go/v17/parquet/pqarrow"
 	"github.com/apache/arrow/go/v17/parquet/schema"
 	libthrift "github.com/apache/thrift/lib/go/thrift"
 	"github.com/stretchr/testify/assert"
@@ -446,3 +449,53 @@ func TestRleBooleanEncodingFileRead(t *testing.T) {
 
 	assert.Equal(t, expected, values[:len(expected)])
 }
+
+func TestDeltaByteArray(t *testing.T) {
+	dir := os.Getenv("PARQUET_TEST_DATA")
+	if dir == "" {
+		t.Skip("no path supplied with PARQUET_TEST_DATA")
+	}
+	require.DirExists(t, dir)
+
+	expected, err := os.ReadFile(path.Join(dir, "delta_byte_array_expect.csv"))
+	require.NoError(t, err)
+	csvReader := csv.NewReader(bytes.NewReader(expected))
+
+	records, err := csvReader.ReadAll()
+	require.NoError(t, err)
+
+	records = records[1:] // skip header
+
+	props := parquet.NewReaderProperties(memory.DefaultAllocator)
+	fileReader, err := file.OpenParquetFile(path.Join(dir, "delta_byte_array.parquet"),
+		false, file.WithReadProps(props))
+	require.NoError(t, err)
+	defer fileReader.Close()
+
+	arrowReader, err := pqarrow.NewFileReader(
+		fileReader,
+		pqarrow.ArrowReadProperties{BatchSize: 1024},
+		memory.DefaultAllocator,
+	)
+	require.NoError(t, err)
+
+	rr, err := arrowReader.GetRecordReader(context.Background(), nil, nil)
+	require.NoError(t, err)
+	defer rr.Release()
+
+	for rr.Next() {
+		rec := rr.Record()
+		defer rec.Release()
+
+		for i := 0; i < int(rec.NumCols()); i++ {
+			vals := rec.Column(i)
+			for j := 0; j < vals.Len(); j++ {
+				if vals.IsNull(j) {
+					require.Equal(t, records[j][i], "")
+					continue
+				}
+				require.Equal(t, records[j][i], vals.ValueStr(j))
+			}
+		}
+	}
+}