[
https://issues.apache.org/jira/browse/PIG-1518?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12903283#action_12903283
]
Ashutosh Chauhan commented on PIG-1518:
---------------------------------------
Yan,
Sorry for being late on this now thats its committed. But I think you have
gotten it other way around. A CollectableLoadFunc is combinable but
OrderedLoadFunc is not. Lets go over all three interfaces:
* h4. CollectableLoadFunc: A loader implementing it must make sure that all
instances of a particular key is present in one split. If you combine splits of
such a loader, it will still remain CollectableLoadFunc because all instances
of keys will still be in same split after combination. It is dictating a
property *within* a split. Thus, its combinable.
* h4. OrderedLoadFunc: OrderedLoadFunc insists that loader implementing it must
read splits in a well defined order. If you combine the splits, that order may
not hold. You cant combine splits for this loader. Its defining a property
*across* multiple splits.
* h4. IndexableLoadFunc: Says that loader is indexable meaning given a key it
will get you as close as possible to that key. It inherently assumes data is
sorted and index is built for it. Your combined splits may not remain sorted
anymore. You cant combine splits for this interface either. Its defining a
property *across* multiple splits.
If you agree with above then PigStorage isnt combinable because
{code}
public class PigStorage extends FileInputLoadFunc implements
StoreFuncInterface, LoadPushDown{}
and
public abstract class FileInputLoadFunc extends LoadFunc implements
OrderedLoadFunc {}
{code}
I also didnt get your logic for *CollectableLoadFunc AND a OrderedLoadFunc* It
will help if you can explain that a bit.
> multi file input format for loaders
> -----------------------------------
>
> Key: PIG-1518
> URL: https://issues.apache.org/jira/browse/PIG-1518
> Project: Pig
> Issue Type: Improvement
> Reporter: Olga Natkovich
> Assignee: Yan Zhou
> Fix For: 0.8.0
>
> Attachments: PIG-1518.patch, PIG-1518.patch, PIG-1518.patch,
> PIG-1518.patch, PIG-1518.patch, PIG-1518.patch, PIG-1518.patch, PIG-1518.patch
>
>
> We frequently run in the situation where Pig needs to deal with small files
> in the input. In this case a separate map is created for each file which
> could be very inefficient.
> It would be greate to have an umbrella input format that can take multiple
> files and use them in a single split. We would like to see this working with
> different data formats if possible.
> There are already a couple of input formats doing similar thing:
> MultifileInputFormat as well as CombinedInputFormat; howevere, neither works
> with ne Hadoop 20 API.
> We at least want to do a feasibility study for Pig 0.8.0.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
