Hi Ranjan, Do you mean the tuples all have 3 fields, with the 2nd and 3rd being chararrays? Please take a look at bincond ( http://pig.apache.org/docs/r0.9.1/basic.html), it might be something you are looking for.
A = load 'input' as (f1, f2, f3); B = foreach A generate (f1 is null? f1 : CONCAT(f2, f3); /** Assuming f2 and f3 are char/bytearrays. Else handle it however you would like to combine) Thanks, Prashant On Mon, Feb 13, 2012 at 10:47 AM, Ranjan Bagchi <[email protected]>wrote: > Hi, > > I'm processing a set of log files and trying to do some simple data > cleaning on them. It's of the form: given a tuple, if value 1 isn't null, > use it, otherwise combine values 2 and 3. This is pretty simple default > stuff. > > Is this even expressable in pig latin? Or do I need to write a UDF? > > Thanks, > > Ranjan
