Norbert Orzechowicz
Norbert Orzechowicz
One of the things that will help to consume/expose Http-based datasets would be to be able to infer/generate OpenAPI specification for a given dataset. Here is an example scenario, let's...
Flow is amazing, but the true power would come with a parallel processing of datasets. The absolute minimum we need to prepare for that is: - parallel data extraction -...
One of the challenges of working with large systems is state reproducibility. In order to fix any bug, test something manually, or simply play with the system locally, we need...
Sometimes when converting from one data format to another some metadata is missing, for example, when converting from PHP int into Parquet int we can't say for sure if it's...
The float entry type is good enough for basic operations. However, it might still lead to floating error since floats are not designed for precision, float should also only allow...
As pointed out in #1322, most of the mathematical operations (comparisons/sub/add, etc) are handled natively by PHP. But since PHP does not come with a Decimal type it can't be...
Currently, all Aggregating Functions are storing aggregated results in memory. This is not fully aligned with Flow core philosophy which says that Flow is supposed to be memory efficient in...
Initially this task was about adding Logger to Flow, but I think we should go beyond that and integrate Flow with OpenTelemetry. Things that I would start from: - measuring...
This is an experimental feature that might not work at all. So the idea is to whenever possible (or when configured) to offload those heavy tasks to better optimized engine....
I noticed that probably not all compoer.json files from flow packages are fully up to date, especially about the required extensions. Those are all extensions the monorepo requires (some might...