Search code examples
hadoophfile

What's the relationship between hadoop's TFile and HFile?


It seems hadoop support both TFile and HFile. I'd like to know the difference between then and how they appeared (e.g. was HFile derived from TFile?).


Solution

  • It seems that http://hbase.apache.org/book.html#_hfile_format_2 has most of answers.

    HFile is derived from TFile. It was designed to improve HBase performance over MapFile format. Sounds like tweaks were made and TFile could not be used as is, so HFile appeared.