Data not appending in dataset through python - outside dss

nadardin
nadardin Registered Posts: 5 ✭✭✭✭

Hi,

I am trying to append the data in my current dataset.

Tried with all method (write_row_dict, write_tuple, write_dataframe).

While executing its shows 1 rows successfully written, but when I go and check dataiku application and check dataset only one row available.

Can any one guide on this ? how to resolve this ?

Thanks,

Answers

  • AlexandreV
    AlexandreV Dataiker, Registered Posts: 28 Dataiker

    Hi @nadardin
    ,

    could you share the code you tried to run?

    What error did you get?

    Alex

  • AndrewM
    AndrewM Dataiker, Dataiku DSS Core Designer, Dataiku DSS Adv Designer Posts: 20 Dataiker
    edited July 17

    Hello,

    If you're using a Python Recipe in the flow you may want to try setting the output dataset to 'Append instead of overwrite' under 'Inputs/Outputs'. This will add new rows to the output dataset every time the recipe is run instead of overwriting the data.

    Screen Shot 2021-01-26 at 2.08.30 PM.png

    You can also use a Pandas data frame append mode in your Python code to append data: https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.append.html. Since we can directly write a dataframe to a dataset in DSS, we can append the new rows to the dataframe in the Python code and then write that dataframe to the dataset. I've included a small sample below that illustrates this. It takes a dataset (test_csv_df) and then appends it to itself in a separate dataframe. The appended data dataframe is then written back to another dataset in DSS.

    # -*- coding: utf-8 -*-
    import dataiku
    import pandas as pd, numpy as np
    from dataiku import pandasutils as pdu
    
    # Read recipe inputs
    test_csv = dataiku.Dataset("Test-csv")
    test_csv_df = test_csv.get_dataframe()
    
    temp_df = test_csv_df.append(test_csv_df) 
    
    # Write recipe outputs
    temp_df = dataiku.Dataset("collapsed-data")
    temp_df.write_with_schema(collapsed_data_df)

    Hope this helps!

    Andrew M

  • nadardin
    nadardin Registered Posts: 5 ✭✭✭✭

    I am trying to append data in input data set. Can we append in input data set

    data = {'xx:'xxx','key':'1212'}

    df = pandas.DataFrame(pandas.json_normalize([data]);

    writer = dataiku.core.Dataset(dataSetName,projectKey);

    vp = dataiku.core.dataset_write.DatasetWriter(writer);

    vp.write_dataframe(df);

    No Error. It's showing 1 rows successfully written

    Thanks

  • AlexandreV
    AlexandreV Dataiker, Registered Posts: 28 Dataiker
    edited July 17

    Hello,
    you could get your output dataset as a dataframe and append the new data to it.

    import dataiku
    import pandas as pd
    
    output_dataset = dataiku.Dataset("out")
    out_df = output_dataset.get_dataframe()
    
    data = [
        {'xx':'row 1','key':'1212'},
        {'xx':'row 2','key':'1212'},
    ]
    df_to_append = pd.json_normalize(data)
    
    with out_df.get_writer() as writer:
        writer.write_dataframe(out_df.append(df_to_append))

    Hope this solves your issue

    Alex

Setup Info
    Tags
      Help me…