In a use case I am working on I need to search by ID's on an alias that has about 100 indices. I randomly picked about 30 ID's and ran both mget and scroll by ids and I get inconsistent timings out of each, which is confusing and not sure which is better solution for use case like this ?
Run
mget
scroll
1
108.213038ms
237.588876ms
2
256.001204ms
161.914637ms
3
202.533423ms
107.588615ms
4
211.439077ms
114.235954ms
Any suggestions or details how each of these gets executed under hood when doing it via alias ?
Thanks
Edit: Just realized that mget doens't work on alias at all, thought it was working because top level query doesn't return error, but individual queries does
How many IDs are in your query ? I'm not certain to understand your problem.
To my mind, if you run an Ids Query, there is no reason to scroll. If you pass 1000 ids, set size to 1000, if there is more than 1024, run a multirequest, or loop over your array?
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.