Update HDo file with hadoop file

Question

I am new to Hadoop. I read that HDFS mainly deals with "write once, read anytime." I have a use case where I may have to make changes to files stored in HDFS. I am exploring if there are ways to do this.

My question is: can I upload an HDFS file to HBase, make changes and save it back to HDFS and delete the original file. Please allow me if possible.

Answer 1

If you need to update the values ​​in a file, you are much better off using HBase. You can still use the HBase table in your MR jobs through TableInputFormat and TableOutputFormat. If you want to add data, you can use any version of hadoop that supports hdfs append, for example 0.20.205.0.