Hi Chris and all,
Just to follow up on this previous discussion. I have reached out to the Apache Druid community with a concrete proposal for adding Apache TsFile input format support to Druid: https://lists.apache.org/thread/qmrkhobp7p07wkmcf4y4myq36k866zz0 The initial feedback from the Druid community is positive. The suggested direction is to start with a Druid community extension under `extensions-contrib`, implemented through a TsFile `InputFormat` and `InputEntityReader`. The Druid side also mentioned that the TsFile community should take responsibility for maintenance and compatibility testing of the extension. They also suggested that, once a small prototype is ready, it should include sample data, usage examples, dependency/license information, and a basic test plan so that the extension can be reviewed properly. I think this is a useful opportunity to expand TsFile's integration with the broader Apache analytics ecosystem. It is also aligned with the direction discussed earlier in this thread. I would like to hear any thoughts or concerns from the TsFile community, especially around the initial scope, compatibility testing, and long-term maintenance of such an extension. If there are no major concerns, I plan to continue with a small prototype and share updates with both the TsFile and Druid communities. Best regards, Colin At 2026-07-16 11:55:24, "ColinLee" <[email protected]> wrote: > >Hi Chris, > >Thanks for pointing this out. >I agree that Apache Druid would be a good project for TsFile to reach out to, >especially since Druid already supports several columnar and analytical data >formats. > >From the TsFile side, I believe this integration should be feasible. A natural >starting point would be to support TsFile as an input format or ingestion >source in Druid. >We can take a closer look at Druid's existing format and extension mechanisms, >and then reach out to the Druid community with a concrete integration proposal. > >I would be happy to participate in this from the TsFile community side. > >Best, >Colin > > > > > > > > > > > > > > > >At 2026-07-16 03:23:51, "Christofer Dutz" <[email protected]> wrote: >>Hi all, >> >>While reviewing the board report of Apache Druid, I noticed this: >> >>Druid has an extensive web of connections with other Apache projects: Calcite >>for SQL planning, Curator and ZooKeeper for coordination, Kafka, Hadoop and >>Iceberg as >>data sources, Avro, ORC, or Parquet as supported data input formats, and >>DataSketches for scalable approximate algorithms. Druid can also be used as a >>data source by Superset. >> >>However, it seems to be missing TSfile, which should probably also be a good >>thing to support. >> >>Possibly worth reaching out to them? (Or did I already recommend that?) >> >>Chris >>
