add more comments

apache · Aug 28, 2024 · f4720f5 · f4720f5
1 parent 1ffc720
commit f4720f5
Show file tree

Hide file tree

Showing 3 changed files with 5 additions and 6 deletions.
diff --git a/c++/include/orc/Writer.hh b/c++/include/orc/Writer.hh
@@ -293,6 +293,8 @@ namespace orc {
 
     /**
      * Set whether the compression block should be aligned to row group boundary.
+     * The boolean type may not be aligned to row group boundary due to the
+     * requirement of the Boolean RLE encoder to pack input bits into bytes
      */
     WriterOptions& setAlignBlockBoundToRowGroup(bool alignBlockBoundToRowGroup);
 

diff --git a/c++/src/ColumnWriter.cc b/c++/src/ColumnWriter.cc
@@ -1271,11 +1271,7 @@ namespace orc {
 
   void StringColumnWriter::finishStreams() {
     ColumnWriter::finishStreams();
-    if (useDictionary) {
-      dictDataEncoder->finishEncode();
-      dictLengthEncoder->finishEncode();
-      dictStream->finishStream();
-    } else {
+    if (!useDictionary) {
       directDataStream->finishStream();
       directLengthEncoder->finishEncode();
     }

diff --git a/c++/src/ColumnWriter.hh b/c++/src/ColumnWriter.hh
@@ -183,7 +183,8 @@ namespace orc {
      * Finalize the encoding and compressing process. This function should be
      * called after all data required for encoding has been added. It ensures
      * that any remaining data is processed and the final state of the streams
-     * is set.
+     * is set. Note: the boolean type may break this spec due to some trailing bits will be written
+     * to the next compression block.
      */
     virtual void finishStreams();