I want to test two mount methods

Sure, but you can only configure one at a time. So these will be two separate tests

yes, I’m testing the two mount methods in the test environment to see which is more suitable for our online environment @User

Ok, cool. Hope you got answers to your questions @User

We tried star-tree index before loading segments but it doesn't look like it created

table config

index map from server

any errors in config file ?

Probably need to enable generation?

We should have a validation tool for the configs?

i think we by choice allow extra configs to stay there, so that if something is removed, we wont fail

I understand that behavior in controller

But would be good to have a standalone tool that flags these

true

TableConfig

{ "code": 400, "error": "Cannot deserialize instance of

java.util.ArrayList

out of START_OBJECT token\n at [Source: UNKNOWN; line: -1, column: -1] (through reference chain: org.apache.pinot.common.config.IndexingConfig[\"starTreeIndexConfigs\"])" }

Hi. I'm trying to follow the getting started instructions but they seem out of date. E.g. TOP doesn't appear to be allowed anymore. Scripts are missing. Any thoughts? Thanks! - Dan

hey Dan, we've started using https://apache-pinot.gitbook.io/apache-pinot-cookbook/getting-started as the documentation. Could you please try the instruction on this one instead?

Ah, can do. Can someone update the 'learn more' link at https://pinot.apache.org/ ?

yeah, will do

updated the links @User. This documentation is still a WIP. Feedback is welcome!

I've hit two issues with pinot-admin.sh. It seems to fail sometimes and not return an error. How can I get errors to be returned from it? One of the issues was I had a extra comma in the json spec during AddTable. Part of the success output is printed out (no errors). Another issue is I'm trying to batch load a 20M row csv (a 2M row csv worked). There is not an error in the output. https://docs.google.com/document/d/1MqoPzDX9oc-QdQE-pDHw-htKya3tlEOPB9_ZyRCcS28/edit#

Hi. I have a few questions. (1) Does anyone have code where Protocol Buffers are used as inputs for ingestion? (2) Has anyone designed an API in front of Pinot that joins pinot with another data source for the API? E.g. I want to have my stats generated in Pinot and then join them entities in a MySQL table and return the results in GraphQL. It seems like I can use Apache Calcite to do this. (3) Is there documentation on useful/common integrations? I see on the gitbook.io that there's support for Thirdeye, Superset, Presto, Powerbi. Anything else? Thanks! - Dan

These do seem like great integration ideas. We don’t have these (afaik), but happy to help you get these working

For 1) this is streaming or batch mode?

2) you can do this with presto Pinot connector

(1) For both. (2) Cool, I'll take a look at Presto.

Also curious how do you like GraphQL for analytical queries?

My impression from a while back was that it was too verbose to perform a simple aggregation

For my use case, GraphQL is mostly for UIs and external API. For analyzing ad stats, a SQL-like interface is easier. For doing mgmt stuff (like creating, updating, listing, etc), GraphQL is pretty nice. GraphQL lets me fetch multiple levels of entities in the same API call. The fields also can be more complex than converting fields to SQL columns. I'll probably end up exposing both a GraphQL and a SQL-like interface at some point.