Pangool 0.70 released

[Pere Ferrera]

Hello,

I'm glad to announce release 0.70 of Pangool with many new good things:

1- Backwards compatibility: It is now possible to use tuple files which have a Schema that differs with the Schema to be used when reading it. This is useful if Schemas have evolved (i.e. new fields have bee added) and makes sense now that Pangool supports nullable fields. Read more in http://pangool.net/userguide/TupleMrBuilder.html section "Using (binary) Tuples as input/output":

2- New typed getters for Tuple, described in: http://pangool.net/userguide/schemas.html

3- New Mutator class for mutating schemas, described in: http://pangool.net/userguide/schemas.html

4- More flexible named outputs, as discussed in the user list, see section "Advanced: default named output": http://pangool.net/userguide/named_outputs.html

In addition we have solved a couple of small problems and provided full null support for TupleTextOutputFormat which was lacking.

We have also revised the whole pangool.net site and added the new section Text I/O in the user guide, which is something that was not properly documented before: http://pangool.net/userguide/text_io.html

From this point on, we will be looking forward to release a 1.0 version of Pangool. We think we are very near to it, and we are very glad that Pangool is becoming a solid and reliable tool for the Hadoop ecosystem, supporting both MR1 and MR2 and incorporating all the features that make sense for really taking advantage of it.

Cheers,

Pere.