Why the logic of ParquetFileWrite is diff from the parquet format

[asfimport]

Hi parquet-mr teams,

When I reading the parquet writer in [ParquetFileWriter](https://github.com/apache/parquet-mr/blob/e06384455567c56d5906fc3a152ab00fd8dfdf33/parquet-hadoop/src/main/java/org/apache/parquet/hadoop/ParquetFileWriter.java#L916),  I find that there is no column metadata behind the each column chunk described in the [Parquet-Format] (https://github.com/apache/parquet-format#file-format) and [thrift definition](https://github.com/apache/parquet-format/blob/54e53e5d7794d383529dd30746378f19a12afd58/src/main/thrift/parquet.thrift#L790) 

 

Reporter: Kun Liu / @liukun4515

_{Note: This issue was originally created as PARQUET-2162. Please see the migration documentation for further details.}

[asfimport]

Raphael Taylor-Davies / @tustvold:
Possibly related https://issues.apache.org/jira/browse/PARQUET-2139