[
https://issues.apache.org/jira/browse/CAMEL-8542?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Claus Ibsen updated CAMEL-8542:
-------------------------------
Estimated Complexity: Moderate (was: Unknown)
> hdfs & hdfs2 components are merging data locally instead of streaming it
> ------------------------------------------------------------------------
>
> Key: CAMEL-8542
> URL: https://issues.apache.org/jira/browse/CAMEL-8542
> Project: Camel
> Issue Type: Improvement
> Components: camel-hdfs
> Affects Versions: 2.15.0
> Reporter: Sergey Zhemzhitsky
>
> Here is the
> [conversation|http://camel.465427.n5.nabble.com/HDFS2-Component-and-NORMAL-FILE-type-td5764655.html]
> CAMEL-4555 introduced an ability to merge files from within a single
> directory.
> The merge operation is done locally, i.e. by means of creating the whole file
> on the local file system (that may be space and time consuming in case of
> multi -gigabyte, -terabyte files).
> # It will be more efficient to stream these files directly from hdfs, for
> example by wrapping them into
> [SequenceInputStream|http://docs.oracle.com/javase/7/docs/api/java/io/SequenceInputStream.html]
> or something like this
> [MapReducePartInputStreamEnumeration|https://github.com/yahoo/Glimmer/blob/master/src/main/java/com/yahoo/glimmer/util/MapReducePartInputStreamEnumeration.java]
> # It will be really great if there will be an ability to switch merging on
> and off by means of an option or parameter.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
