I just spent some time combing through Arrow JIRA issues that mention "parquet"
We now have 60 Python-related issues appropriately labeled https://cwiki.apache.org/confluence/display/ARROW/Python+Parquet+Development I noted there are some bugs reported that are duplicates of each other, but will need to examine more closely to confirm There's another 17 that are more C++-related https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=95655294 On Mon, Nov 12, 2018 at 1:16 PM Wes McKinney <wesmck...@gmail.com> wrote: > > hi folks, > > As some of you may have noticed, we are accumulating a mountain of > Parquet-related JIRA issues, many of them resulting from people using > Apache Arrow to do data engineering in Python and running into > problems. > > To help with having better visibility into all the relevant Parquet > issues, and with the monorepo merge behind us, I created a couple wiki > pages linked to from the main > https://cwiki.apache.org/confluence/display/ARROW page: > > * C++ issue dashboard: https://cwiki.apache.org/confluence/x/fpWzBQ > * Python issue dashboard: > https://cwiki.apache.org/confluence/display/ARROW/Python+Parquet+Development > > Many Parquet issues in the ARROW project are not found in these > dashboards because they lack the "parquet" label. Please help with > project organization by remembering to apply the "parquet" label to > any issue. > > Since Ruby also supports Parquet now via GLib, and R support for > Parquet is coming soon, we need to do what we can to grow the > community of people working on the core Parquet libraries and the > things they depend on, like the IO and memory management subsystems of > the Arrow C++ libraries. > > In general, I think it is very important for us to have fast and > reliable C++ support (and language bindings) for the 5 major file > formats in use in data warehousing: > > * CSV > * JSON > * Parquet > * Avro > * ORC > > Antoine has been leading efforts on reading CSV files, and we will > need to make a push into JSON and Avro at some point. > > Thanks > Wes
