broadinstitute / gatk-dataflow

Development dataflow
BSD 3-Clause "New" or "Revised" License
4 stars 1 forks source link

Add mechanism to generate unique ids for data types (reads, variants, etc.) #10

Open akiezun opened 9 years ago

akiezun commented 9 years ago

From @droazen on May 28, 2015 18:28

Needed for GroupByKey, since Java serialization is not deterministic. Initial idea is to create IDs based on the source of each record (eg., URI + file offset or record number).

Copied from original issue: broadinstitute/hellbender#532