[GitHub] nifi issue #2666: NIFI-5130 ExecuteInfluxDBQuery processor chunking support
Github user michalmisiewicz commented on the issue: https://github.com/apache/nifi/pull/2666 @MikeThomsen I've updated the code. Please let me know if I should change maximum chunk size. ---
[GitHub] nifi issue #2666: NIFI-5130 ExecuteInfluxDBQuery processor chunking support
Github user michalmisiewicz commented on the issue: https://github.com/apache/nifi/pull/2666 @MikeThomsen it supposed to fetch the entire result set of that select statement each run ---
[GitHub] nifi issue #2666: NIFI-5130 ExecuteInfluxDBQuery processor chunking support
Github user MikeThomsen commented on the issue: https://github.com/apache/nifi/pull/2666 @michaelandrepearce so I pushed about 30k events into InfluxDB using some artifacts from @mans2singh's PRs and ran `select * from /*/` on the database. Let it run twice and the result sets were different in size. Is each iteration of the processor supposed to be a new chunk or is it supposed to chunk the entire result set of that select statement each run? ---
[GitHub] nifi issue #2666: NIFI-5130 ExecuteInfluxDBQuery processor chunking support
Github user MikeThomsen commented on the issue: https://github.com/apache/nifi/pull/2666 @mans2singh do you have free time to help review this? ---
[GitHub] nifi issue #2666: NIFI-5130 ExecuteInfluxDBQuery processor chunking support
Github user michalmisiewicz commented on the issue: https://github.com/apache/nifi/pull/2666 @MikeThomsen yes. Chunking is the only way to fetch unlimited number of rows in a single query from InfluxDB that has set `max-row-limit` property. > Many production InfluxDB installation has limited number of rows returned in a single query (by default 10k). Chunking query can return an unlimited number of rows. ---
[GitHub] nifi issue #2666: NIFI-5130 ExecuteInfluxDBQuery processor chunking support
Github user MikeThomsen commented on the issue: https://github.com/apache/nifi/pull/2666 @michalmisiewicz so if I understand the Jira ticket correctly, the purpose of this ticket is to make it possible to stream out everything that matches the query, not just one large result payload. Is that right? ---