hi,
i have logs in this format and i want to start the aggregation when start comes in the line and end the aggregation when end occurs. the aggregation is based on "username".
i was able to use aggregate function but it does not split the documents.
username;command;start/end
user1;command1;
user1;command2;start
user2;command3;start
user3;command4;
user1;command5;
user2;command6;
user3;command7;start
user2;command8;end
user1;command9;
user2;command10;end
user3;command11;end
user2;command12;
user1;command13;end
user2;command14;
user1;command15;
user3;command16;
You want command12 and command14 aggregated even though there is no start/end, but you do not want command1 aggregated? So until a start appears for a user you do not want aggregation, but once a start has been seen you want everything aggregated?
Hi,
Sorry my bad, those two should be separate documents.
Yes, all documents should be individually captured. Once start starts aggregation should happen and aggregation should end once end is found.
Post end till new start per user no aggregation.
I tried doing this with three aggregate filters (similar to example 1 in the documentation) and could not get it to work. I then decided to do it in ruby.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.