Option to set the "ID" of a mash sketch

[tseemann]

I want my IDs to be neat, but mash is making it hard?

Case 1:

mash sketch -r -o foo DIR/R1.fq.gz
mash info foo.msh

  [Hashes]  [Length]  [ID]            [Comment]
  1000      105082    DIR/R1.fq.gz  - 

Case 2 (see closed #32)

cat R1.fq.gz R2.fq.gz  | mash sketch -r -o foo -
mash info foo.msh

 [Hashes]  [Length]  [ID]  [Comment]
  1000      1253918   -      -

I would just like a -n or -I option to set the [ID].

And the [Comment] ?

[ondovb]

Implemented as -I and -C (comment).

[tseemann]

@ondovb I can't see any commits in master that refer to this?

Any chance of a 2.1 release for this feature?

Currently doing crazy named pipe setups to get proper IDs because when I do <(zcat R1 R2) i get the same name on everything /dev/fd/63.

[tseemann]

UPDATE: i see it's in the triangle mod. Triangle does not work with the tools I have. It will need to be a full matrix, sorry.

[ondovb]

It's in 03b28ed, which addresses both the ID issue and the read pair merging (I labelled round robin since it takes turns pulling reads from the input files to preserve -c, which I don't think anyone uses anyway :P ).

It's only part of the sketch command so it shouldn't matter whether you use dist or triangle
following that. Setting IDs directly in those would be possible but a little trickier parameter-wise.

Hopefully I'll get a release out soon but I want to make sure it's stable since there are some sketching changes. If you manage to break something in tip that would be helpful :).

[ondovb]

Ok, published v2.1.

[tseemann]

Thanks @ondovb - I am going to run it now on 25,000 isolates. Will let you know what happens.

[ondovb]

This seems fixed? Reopen if not.