PJ Van Aeken created FLUME-2281:
-----------------------------------
Summary: File Header available in Deserializer
Key: FLUME-2281
URL: https://issues.apache.org/jira/browse/FLUME-2281
Project: Flume
Issue Type: Improvement
Reporter: PJ Van Aeken
At the moment, the file header in a Spooling Directory source is set after the
event is deserialized. So you cannot use the file header to make more
intelligent deserialization decisions based on the type of file you are dealing
with.
A possible solution would be to simply move the code that sets the header. In
the org.apache.flume.client.avro.ReliableSpoolingFileEventReader, you can move
this section (line 254 - 259) :
if (annotateFileName) {
String filename = currentFile.get().getFile().getAbsolutePath();
for (Event event : events) {
event.getHeaders().put(fileNameHeader, filename);
}
}
And then move it to line 239, right above the deserialization code.
--
This message was sent by Atlassian JIRA
(v6.1.4#6159)
