mito
mito copied to clipboard
Dates in column headers errors
Describe the bug Mito does a bad job of handling dates in column headers.
To Reproduce Run this code in a Jupter notebook
import pandas as pd
import numpy as np
import mitosheet
zz = np.random.rand(10000,10000)
def create_header():
names = []
for i in range(10000):
names.append(f'Path_{i}')
return names
col_head = create_header()
infer_datetime_format=True
z = pd.date_range(start='18/12/1994', end='05/04/2022')
df =pd.DataFrame(zz, columns = col_head)
df.insert(loc=0, column='Date', value = z)
df.set_index('Date', inplace= True)
df1 = df.T
df1.head(5)
mitosheet.sheet(df1, view_df=True)
![Screen Shot 2022-04-05 at 10 29 43 AM](https://user-images.githubusercontent.com/18709905/161777137-17edaec6-019c-482f-9649-b13dc9b726bc.png)
Ok, the update is I spent an hour on this and I was not able to create a statisfactory solution. The main issue is that the only two solution I can think of that seems straight forward seem to result in a performance impact that we don't want to go for.
If you want to see it, check out the branch date-header-fix
. Check out the two different commits - they each try two solutions. Even on running the tests alone, they are noticeable (about 30%) slower. Maybe this was just a fluke, so try yourself.
Let me know thoughts - feel free to take a shot.