I just did some test and seems it works for me. How does your define statement look like? How about setting "mapred.cache.files"?
BTW, if you only use UDF not streaming, you can override getCacheFiles in EvalFun starting 0.9. Daniel On Fri, Sep 2, 2011 at 10:16 AM, Deepak Reddy <[email protected]> wrote: > Hi, > > I am trying to use a cached filed called GeoLiteCity.dat.gz#datafile in my > pig script. > > For that I used the CACHE keyword as CACHE('HDFS archivefile#symlink'); > > But when I try to refer to this file in my UDF its not really working. > > So I tried the work around as mapred.cache.archives HDFS archive#symlink and > mapred.create.symlink as yes > > Even this didn't work, but when I created a unzipped file and used it as part > of mapred.cache.archives the UDF returned the right results. > > The same didn't work when I used CACHE('HDFS unzippedfile#symlink'); > > Can you please tell me if these are known issues in pig 0.8 or am I doing > something wrong. > > Thank you, > Deepak Reddy > > >
