Hamilton Open Source

image.png

<@U06J7KZEWBD> what’s the last line of the stacktrace?

```ShapeError: could not create a new DataFrame: series "full" has length 1561990 while series "removed" has length 84846```


For some reason, its trying  to do a pl.DataFrame on both outputs

what do you want result to be?
A dict with two keys? or?

Exactly what is shown on the dag, 3 diferent dataframe.

yep — what is the adapter that you’re creating?

yep. So — what that is doing is trying to “reduce” all the outputs you specified, into a single polars dataframe.

so if you just change things to this:
```from hamilton import base
hamilton_driver = driver.Driver(config, wind_ml.h_functions, adapter=base.DefaultAdapter())```
`result` will now be a dictionary with what you expect.

So the TL;DR: is, the output type of `execute()` can be changed based on the adapter you provide.

Hmm ok. But why would it reduce into one if I used the extract fields decorator to get the values from the dict?

&gt;  But why would it reduce into one
that’s a function of the adapter provided. So if you provide a “polars dataframe result” one, it’ll try to reduce all the outputs into one dataframe.

So I can't ask more than 1 result of the same type?

Yep — did you try running it with:
```from hamilton import base
hamilton_driver = driver.Driver(config, wind_ml.h_functions, adapter=base.DefaultAdapter())```
`base.DefaultAdapter()` here “reduces” results into a dictionary.