9/n remove deprecated ColumnChunk.file_offset

apache · Sep 2, 2024 · f273033 · f273033
1 parent e5eff65
commit f273033
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 18 deletions.
diff --git a/cpp/src/generated/parquet3_generated.h b/cpp/src/generated/parquet3_generated.h
diff --git a/cpp/src/parquet/metadata3_benchmark.cc b/cpp/src/parquet/metadata3_benchmark.cc
@@ -116,6 +116,17 @@
 // 5/large-footer2: num-rgs=4 num-cols=2930 thrift=2248476 flatbuf=1165112
 //
 //
+// Remove deprecated ColumnChunk.file_offset
+//
+//
+// 0/amazon_apparel.footer: num-rgs=1182 num-cols=16 thrift=2158995 flatbuf=1292376
+// 1/amazon_movie_tv.footer: num-rgs=3 num-cols=18 thrift=22578 flatbuf=5056
+// 2/amazon_polarity.footer: num-rgs=900 num-cols=4 thrift=1074313 flatbuf=214192
+// 3/amazon_reviews_books.footer: num-rgs=159 num-cols=44 thrift=767840 flatbuf=226112
+// 4/large-footer1: num-rgs=23 num-cols=2001 thrift=3253741 flatbuf=2961808
+// 5/large-footer2: num-rgs=4 num-cols=2930 thrift=2248476 flatbuf=1120360
+//
+//
 
 #pragma GCC diagnostic ignored "-Wunused-parameter"
 #pragma GCC diagnostic ignored "-Wunused-function"
@@ -375,11 +386,7 @@ class Converter {
   auto operator()(format::ColumnChunk& cc, int rg_idx, int col_idx) {
     auto meta_data = (*this)(cc.meta_data, rg_idx, col_idx);
 
-    // All offsets are relative to the row group.
-    auto off = md_->row_groups[rg_idx].file_offset;
-
     format3::ColumnChunkBuilder b(builder_);
-    b.add_file_offset(cc.file_offset - off);
     b.add_meta_data(meta_data);
     cc.__isset.crypto_metadata = false;  // TODO
     if (cc.__isset.encrypted_column_metadata) {

diff --git a/cpp/src/parquet/parquet3.fbs b/cpp/src/parquet/parquet3.fbs
@@ -174,7 +174,6 @@ table ColumnMetadata {
 
 table ColumnChunk {
   file_path: string;
-  file_offset: int;
   meta_data: ColumnMetadata;
   // crypto_metadata: ColumnCryptoMetadata;  // TODO
   // encrypted_column_metadata: [byte];  // TODO