0

I want to override Hadoop's default record reader within my Java code. It reads records line-by-line by default. I changed it when it encounters ":", separates the records. However, while seprating the records, it also emits the delimiter. So for example, my records are:

> I go to school : He goes to school : we go to school

When records are read, it produces like:

I go to school
He goes to school
we go to school

However the delimiter - ":" is gone. Is there way to keep this delimiter in records? Because there may be some important informations in delimiters.

Jeyhun Karimov
  • 1,295
  • 19
  • 27
  • What kind of important information? Generally, the whole point of a delimiter is that it sits in-between the important information. Can you show us your current record reader code? – DNA Jun 26 '14 at 21:22
  • Thank you for you reply, for example delimiteer maybe some date. Or the ID of the record. In my case the delimiter is date of transactions. For example: "19.11.2014 11:44 Transaction begin: blabla 19.12.2014 11.22 Transaction failed blabla" – Jeyhun Karimov Jun 26 '14 at 21:47

0 Answers0