How to create a dask time series data frame with multiple columns - dask

Question

I'm having trouble creating a dask time series dataframe that calculates the mean per hour over multiple columns.
This is an example of my input csv file:
name,date_time,num
dan,2019-01-02 00:00:00,3
ben,2019-01-02 00:00:00,7
dan,2019-01-02 02:00:00,13
dan,2019-01-02 10:00:00,9
dan,2019-01-02 10:01:00,3
ben,2019-01-02 14:22:00,66
ben,2019-01-02 14:37:00,37
I can produce the desired output using pandas
import pandas as pd
from matplotlib import pyplot
df = pd.read_csv('my_file.csv')
df['timestamp'] = pd.to_datetime(df.date_time)
df = df.set_index(df.timestamp) # set a datetime index
df = df.groupby('name').resample('H')['num'].mean().unstack('name')
df.fillna(0).plot()
Desired output
name ben dan
timestamp
2019-01-02 00:00:00 7.0 3.0
2019-01-02 01:00:00 NaN NaN
2019-01-02 02:00:00 NaN 13.0
2019-01-02 03:00:00 NaN NaN
2019-01-02 04:00:00 NaN NaN
2019-01-02 05:00:00 NaN NaN
2019-01-02 06:00:00 NaN NaN
2019-01-02 07:00:00 NaN NaN
2019-01-02 08:00:00 NaN NaN
2019-01-02 09:00:00 NaN NaN
2019-01-02 10:00:00 NaN 6.0
2019-01-02 11:00:00 NaN NaN
2019-01-02 12:00:00 NaN NaN
2019-01-02 13:00:00 NaN NaN
2019-01-02 14:00:00 51.5 NaN
My attempt to produce the same dataframe with dask
from dask import dataframe as dd
from matplotlib import pyplot
ddf = dd.read_csv('my_file.csv')
# setting an index
ddf['timestamp'] = dd.to_datetime(ddf.date_time)
ddf = ddf.set_index(ddf.timestamp)
ddf.repartition(freq='MS')
ddf.groupby('name').resample('H')['num'].mean()
When I run the code above I get this error:
AttributeError: 'Column not found: resample'
This has me really stumped and any help would be appreciated.

Answer 1

It looks like dask dataframe does not implement a groupby-resample operation. It sounds like you have a feature request. I recommend raising an issue at https://github.com/dask/dask/issues/new
See https://docs.dask.org/en/latest/support.html#asking-for-help for requests on where to ask for help.