Dear Prakhar,

Hope you are doing well.

Please accept our sincerely apology for delay in response.

Hadoop framwork can compress sequence file at the record level or the block level.
The key io.seqfile.compression.type controls the record or block-level-compression for sequence files.A value of BLOCK requests block-level-compression.A value of RECORD, the default, specifies record-level-compression.
A value of NONE disables compression.

In general block-level-compression is recommended, because it provides greater data compression.The compression overhead is less, and the compression ration is much greater.For sequence files that are being used as input to a map or reduce phase, block-level-compression is ideal.

You can enable compression in hdfs-site.xml as follows


Hope it resolves your query.

If you have any further query,please let us know.