[agg-sum] sum() takes no keyword arguments

[geekscrapy]

Small description
sum aggregator no longer working. when applied to a column and a freq sheet is opened you get the following error in the sum agg col: sum() takes no keyword arguments

Expected result
sum aggregation on columns

Actual result with screenshot

Traceback (most recent call last):
  File "/Users/geekscrapy/envs/py37/lib/python3.7/site-packages/visidata/threads.py", line 200, in _toplevelTryFunc
    t.status = func(*args, **kwargs)
  File "/Users/geekscrapy/envs/py37/lib/python3.7/site-packages/visidata/aggregators.py", line 183, in memo_aggregate
    dispval = col.format(typedval)
  File "/Users/geekscrapy/envs/py37/lib/python3.7/site-packages/visidata/column.py", line 229, in format
    return self.make_formatter()(*args, **kwargs)
  File "/Users/geekscrapy/envs/py37/lib/python3.7/site-packages/visidata/column.py", line 209, in _format_len
    return self.formatValue(typedval, **kwargs)
  File "/Users/geekscrapy/envs/py37/lib/python3.7/site-packages/visidata/column.py", line 244, in formatValue
    return vd.getType(self.type).formatter(self.fmtstr, typedval)
  File "/Users/geekscrapy/envs/py37/lib/python3.7/site-packages/visidata/type_floatsi.py", line 10, in SIFormatter
    while abs(val) > 1000:
TypeError: bad operand type for abs(): 'TypedWrapper'

Steps to reproduce with sample data and a .vd

echo -e "A,B,C\n1,2,3" | vd -f csv -N
z+, sum

[anjakefala]

Oh, thank you!

It is related to this: #2009 which is related to 1c466d8. I will take a look soon.

[anjakefala]

@geekscrapy, I am a bit confused. =(

Are you sure about this stacktrace? Were the values typed as type-floatsi before summed? Which commit are you on? Could you give me a gif or series of screenshots?

When I try it on 9ec71f3 I cannot reproduce. I also cannot with #2009. I am using type-int. I also cannot reproduce with type-floatsi.

[geekscrapy]

Are you sure about this stacktrace? Were the values typed as type-floatsi before summed?

Yes, they were.

Tbh, I thought I was on the latest commit but it appears my update func failed. I'll update again and retest

[geekscrapy]

Just realised why it didn't update, you are no longer supporting 3.7 °-° !

Yup, this now works... Sorry, hadn't noticed the version req

[anjakefala]

Oh wow!

Was it a big deal for you to update Python versions?

[geekscrapy]

Massive, I have a massive amount of cli tools that work with different versions of python, and some silently break of not setup correctly.
In the end I just created a new VM and retested.
But for now I've had to roll back the VD version to 4ffacbb on my main host, and it looks like I'll be stuck there for a while as the tool I use vd within doesn't take well to updates 🥴
Hopefully building a new system soon where I can sort out the python envs again, it's a cyclical process 🤷‍♂️

(FWIW, I love this commit, a few great features and feels solid)

[anjakefala]

Ah, I'm sorry it caught you by surprise! =( I tried to advertise this in a few places: #2004

FWIW, we are not making any grand migration to Python 3.8 exclusive code, other than sum(start=). So if you are comfortable with git, you can grab newer stuff that you like. I will keep this in mind when you submit bug reports!

[anjakefala]

Hey @geekscrapy!

We really value all of your contributions to VisiData, so we reintroduced Python 3.7 support.

I would still get yourself upgraded as soon as possible for a bundle of reasons!

[geekscrapy]

Ha ha, thanks! I shall update (vd) immediately 😏 but yes, in going to skip to py12 and be done with it for a year or so !

[anjakefala]

@geekscrapy LMK when you do. =)