adding comments

parquet-hadoop/src/main/java/parquet/hadoop/ParquetFileReader.java

diff --git a/parquet-hadoop/src/main/java/parquet/hadoop/ParquetFileReader.java b/parquet-hadoop/src/main/java/parquet/hadoop/ParquetFileReader.java
index 71a8f74..58d2a8d 100644
--- a/parquet-hadoop/src/main/java/parquet/hadoop/ParquetFileReader.java
+++ b/parquet-hadoop/src/main/java/parquet/hadoop/ParquetFileReader.java
@@ -512,8 +512,14 @@
       try {
         pageHeader = Util.readPageHeader(this);
       } catch (IOException e) {
-        this.pos = initialPos;
-        pageHeader = Util.readPageHeader(new SequenceInputStream(this, f));
+        // this is to workaround a bug where the compressedLength
+        // of the chunk is missing the size of the header of the dictionary
+        // to allow reading older files (using dictionary) we need this.
+        // usually 13 to 19 bytes are missing
+        // if the last page is smaller than this, the page header itself is truncated in the buffer.
+        this.pos = initialPos; // resetting the buffer to the position before we got the error
+        LOG.info("completing the column chunk to read the page header");
+        pageHeader = Util.readPageHeader(new SequenceInputStream(this, f)); // trying again from the buffer + remainder of the stream.
       }
       return pageHeader;
     }