summarise not working

[AInisha]

datar version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of datar and its backends.

Issue Description

I am unable to use summarise(CNT= n()), i donot see a separate column with CN

dataframe= dataframe>>group_by(f.some_col1, f.some_col2)>>summarise(CNT= n())>>ungroup() >> arrange(f.some_col2)

dataframe = (dataframe.set_index('some_col2')
                                .groupby('some_col1', sort=False)
                                .resample('MS')['CNT'].max()
                                .unstack()
                               )

error throws at .resample('MS')['CNT'].max(), as there is no CNT column exist, but earlier project ran fine for months.
Please assist me as early as possible on this
tried below 2 version as these are compatible with my project, earlier it was running fine:-

pip install -U datar==0.9.1
pip install -U datar[pdtypes]==0.9.1

pip install -U datar==0.9.0
pip install -U datar[pdtypes]==0.9.0

Expected Behavior

the column CNT existed before

Installed Versions

Replace this line with the output of datar.get_versions()
0.9.1

[pwwang]

Please:

1. try the latest version
2. provide a runnable example

[AInisha]

for the latest version i am not bale to use the datar at all

!pip install -U datar==0.12.2

!pip install -U datar[pdtypes]==0.12.2

data = {'Name': ['Tom', 'nick', 'krish', 'jack', 'jack'],
'Age': [20, 21, 19, 18, 16]}

Create DataFrame

df = pd.DataFrame(data)
df>>distinct(f.Name)

below is the error
ContextError: Cannot evaluate ReferenceAttr object without a context.

[AInisha]

our project is built with

pip install -U datar==0.9.1
pip install -U datar[pdtypes]==0.9.1

now i used below:-

# !pip install -U datar==0.9.1
# !pip install -U datar[pdtypes]==0.9.1

data = {'Name': ['Tom', 'nick', 'krish', 'jack', 'jack'],
        'Age': [20, 21, 19, 18, 16]}
  
# Create DataFrame
df = pd.DataFrame(data)
df>>distinct(f.Name)
df>> group_by(f.Name)>>summarise(CNT= n())>>filter(CNT>0)

i see NameError: name 'CNT' is not defined

[AInisha]

there is no count column displays as separate, my point is that, earlier it used to do that

[AInisha]

is there any update

[pwwang]

Please follow the instructions above.

Note that the error you got NameError: name 'CNT' is not defined is because that you need f.CNT.

[AInisha]

let me try it with libraries you have used

[AInisha]

still getting similar errors:
I am having below as the output from get_versions, i think its due to version compatibility , Thank you for your help
python : 3.8.10
datar : 0.12.2
simplug : 0.3.2
executing : 1.2.0
pipda : 0.12.0
datar-numpy : 0.2.0
numpy : 1.24.4
datar-pandas: 0.3.0
pandas : 2.0.3

[pwwang]

Can you attach the full snapshot like I did, including the code and the error?

[pwwang]

I mean the full snapshot please, including the version printing, and module importing.

[AInisha]

I ran without downloading any additional library and exceuted only latest datar and imported its library, it ran fine gives separate column of count, showing the pandas version like -->pandas : 1.2.4, python 3.8, other version remained same like your screenshot.
issue can be because our pandas version is -->2.0.3, when executing with libraries, some libraries which we need automatically takes pandas 2.0.3 version i guess, is there any restriction to pandas version?

[pwwang]

Yes, pandas needs to be 1.x, pandas 2.x is a different story.

[AInisha]

Will there be any fix by this week so that datar latest version works with pandas >= 2.0.3

[pwwang]

No, pandas 2 uses pyarrow as backend. Quite a lot of stuff need to be reimplemented. But it's in the plan.

[GitHunter0]

Will there be any fix by this week so that datar latest version works with pandas >= 2.0.3

@pwwang is incredibly efficient, but it's hard to keep up with so many changes.
dplyr 1.1.0, for example, received a lot of updates too.

[pwwang]

Thanks for bringing this. We are now aligning with dplyr 1.0.8