Python Tool throws RuntimeError on Alteryx.read() if there are no rows

The Python Tool is throwing a RuntimeError if a input table to the Alteryx.read() function contains no rows.

The error shown is:

Error: unable to read input table "4460abb7be83bae8f01b9bf1238a923c" (C:\Users\...

I've created a workaround by inputting the column names separately, catching the error, and exporting an empty Pandas DataFrame, but the Alteryx.read() function should handle tables with no rows by creating an empty Pandas DataFrame directly.

The implementation is part of a Macro and I can't guarantee they will always be rows received.

Had anyone else encountered this problem?

Error Message

Python

Macros

Accepted answers

All comments

WilliamR

Hello @braggis ,

I experience an error when no data is passed to the python tool.

As another workaround, you can add a record count with the RecordCount tool with an AppendField to be sure that you push every time a data frame with values inside.

(If this post helps, then please consider Accept it as the solution to help the other members find it more quickly).

SeanAdams

Hi @WilliamR

cc: @BlytheE

I battled with this exact error today too!

Is there a fix on the roadmap for this? It seems like the desired behavior would be for the Alteryx.Read to return a dataframe with columns but no rows rather than throwing an error that cannot be managed?

it is possible to put the Alteryx.read('#1') into a try-except block - but if it throws an error you just don't know if the error was caused by a missing connection; a failure of the Alteryx library or an empty record-set (which should not be a terminal error condition)! Doesn't make sense for every user to have to work around this by putting a count tool in front of every python tool just to make sure it doesn't fail because of an empty recordset coming into the input.

If it helps - I'm happy to log an idea for this?

Thanks @WilliamR

Sean

Thableaus

Hey Alteryx team, no updates on this? I agree with @SeanAdams it makes no sense at all

jineshnp31

Hi, checking to see if there was any update on this? Facing the same issue. A fix for this will be great! Thanks.

Felipe_Ribeir0

Hi, just to share the simple workaround that is working for me until this is fixed, i am uniting a text input with null data and filtering out this row inside the python tool.

TimPowers

Checking on this myself.. At least for my use-case there will be some data somewhere, so I'll just use try/except to concatenate incoming lines.

V_krishnamoorthy4

BELOW IS THE ANSWER

#data = Alteryx.read("#1")

# Define the file path (you can also pass this as an input parameter)
file_path = r"yourfile.csv"

# Read only the column headers
try:
df = pd.read_csv(file_path, nrows=0)
column_headers = df.columns.tolist()
print(column_headers)

# Output the column headers as a DataFrame
output_df = pd.DataFrame(columns=column_headers)
Alteryx.write(output_df, 1)

except FileNotFoundError:
print(f"The file at {file_path} was not found.")
Alteryx.write(pd.DataFrame(), 1)

TimPowers

Yup,

That's basically what I eventually came to do. The Python tool will read in an empty dataframe, so if emptiness is a concern I mark each relevant code block with if len(df.index) > 0: (a try-block may be more conventional but it's wordier), otherwise nothing happens and I build an outgoing empty dataframe, using appropriate headers and data types for the outgoing write. Often use Control Containers with this paradigm now (which can't be used along a Detour branch btw).

Quick Links

This months top contributors

mceleavey 383

mbarone 337

Hollingsworth 335

LanisC 335

RithiS 280