Hi,
When play around PyArrow scalar, I found it seems to expect the input as a
"pure Python object", e.g. it cannot be a Python list of arrow scalar (such
as `[ pa.scalar(1) ]`:
```
>>> import pyarrow as pa
>>> pa.__version__
'7.0.0'
>>> pa.scalar([1])
>>> pa.scalar([pa.scalar(1)])
Traceback
Yes and no :) Disclaimer: this answer is a little out of my
wheelhouse as I've learned relational algebra through doing and so my
formal theory may be off. Anyone is welcome to correct me. Also,
this answer turned into a bit of ramble and is a bit scattershot. You
may already be very familiar
Hi,
Is there a way to evaluate mathematical expressions against columns of a
pyarrow table which is in memory already similar to how projections work
for dataset scanners?
The goal is to have specify a bunch of strings like sum(a * b)/sum(a), or
avg(a[:10]) etc. Convert these into expressions