ValueError: Metadata mismatch: Expected partition of type `DataFrame` but got `DataFrame`

I’m getting an error which suggests a possible dask validation code issue:

ValueError: Metadata mismatch found in from_delayed. Expected partition of type DataFrame but got DataFrame

Since the expected type is the actual type and it is still throwing an error, it leads me to believe that perhaps the dask validation routine is not working well.

This happens when trying to combine several futures for csv loads. The code to reproduce this error is as follows:

from dask.distributed import Client, progress, get_worker
import dask.dataframe as dd
from dask import delayed

client = Client(<scheduling server host:port>)
load_tasks={}
task_worker_map={"yellow_tripdata_2009-01.csv":"<worker 1 server host:port>",
                 "yellow_tripdata_2017-01.csv":"<worker 1 server host:port>",
                 "yellow_tripdata_2017-02.csv":"<worker 2 server host:port>",   
                }

for file_to_load, worker_to_do in task_worker_map.items():
    fut = client.submit(dd.read_csv,f"/home/saif/{file_to_load}", workers={worker_to_do})
    load_tasks[file_to_load]={"file": file_to_load,
                   "submitted-to": worker_to_do,
                   "future": fut
                  }

futurez = [load_tasks[atask]["future"] for atask in load_tasks.keys()]
df_overarching = dd.from_delayed([delayed(f) for f in futurez])
df_overarching.head()

The full stack trace is:

> 
> -----------------------
> ValueErrorTraceback (most recent call last)
> <ipython-input-17-59d0937d69ae> in <module>
> ----> 1 df_overarching.head()
> 
> ~/miniconda3/envs/churn/lib/python3.6/site-packages/dask/dataframe/core.py in head(self, n, npartitions, compute)
>     898 
>     899         if compute:
> --> 900             result = result.compute()
>     901         return result
>     902 
> 
> ~/miniconda3/envs/churn/lib/python3.6/site-packages/dask/base.py in compute(self, **kwargs)
>     154         dask.base.compute
>     155         """
> --> 156         (result,) = compute(self, traverse=False, **kwargs)
>     157         return result
>     158 
> 
> ~/miniconda3/envs/churn/lib/python3.6/site-packages/dask/base.py in compute(*args, **kwargs)
>     396     keys = [x.__dask_keys__() for x in collections]
>     397     postcomputes = [x.__dask_postcompute__() for x in collections]
> --> 398     results = schedule(dsk, keys, **kwargs)
>     399     return repack([f(r, *a) for r, (f, a) in zip(results, postcomputes)])
>     400 
> 
> ~/miniconda3/envs/churn/lib/python3.6/site-packages/distributed/client.py in get(self, dsk, keys, restrictions, loose_restrictions, resources, sync, asynchronous, direct, retries, priority, fifo_timeout, actors, **kwargs)
>    2566                     should_rejoin = False
>    2567             try:
> -> 2568                 results = self.gather(packed, asynchronous=asynchronous, direct=direct)
>    2569             finally:
>    2570                 for f in futures.values():
> 
> ~/miniconda3/envs/churn/lib/python3.6/site-packages/distributed/client.py in gather(self, futures, errors, maxsize, direct, asynchronous)
>    1820                 direct=direct,
>    1821                 local_worker=local_worker,
> -> 1822                 asynchronous=asynchronous,
>    1823             )
>    1824 
> 
> ~/miniconda3/envs/churn/lib/python3.6/site-packages/distributed/client.py in sync(self, func, *args, **kwargs)
>     751             return future
>     752         else:
> --> 753             return sync(self.loop, func, *args, **kwargs)
>     754 
>     755     def __repr__(self):
> 
> ~/miniconda3/envs/churn/lib/python3.6/site-packages/distributed/utils.py in sync(loop, func, *args, **kwargs)
>     329             e.wait(10)
>     330     if error[0]:
> --> 331         six.reraise(*error[0])
>     332     else:
>     333         return result[0]
> 
> ~/miniconda3/envs/churn/lib/python3.6/site-packages/six.py in reraise(tp, value, tb)
>     691             if value.__traceback__ is not tb:
>     692                 raise value.with_traceback(tb)
> --> 693             raise value
>     694         finally:
>     695             value = None
> 
> ~/miniconda3/envs/churn/lib/python3.6/site-packages/distributed/utils.py in f()
>     314             if timeout is not None:
>     315                 future = gen.with_timeout(timedelta(seconds=timeout), future)
> --> 316             result[0] = yield future
>     317         except Exception as exc:
>     318             error[0] = sys.exc_info()
> 
> ~/miniconda3/envs/churn/lib/python3.6/site-packages/tornado/gen.py in run(self)
>     727 
>     728                     try:
> --> 729                         value = future.result()
>     730                     except Exception:
>     731                         exc_info = sys.exc_info()
> 
> ~/miniconda3/envs/churn/lib/python3.6/site-packages/tornado/gen.py in run(self)
>     734                     if exc_info is not None:
>     735                         try:
> --> 736                             yielded = self.gen.throw(*exc_info)  # type: ignore
>     737                         finally:
>     738                             # Break up a reference to itself
> 
> ~/miniconda3/envs/churn/lib/python3.6/site-packages/distributed/client.py in _gather(self, futures, errors, direct, local_worker)
>    1651                             six.reraise(CancelledError, CancelledError(key), None)
>    1652                         else:
> -> 1653                             six.reraise(type(exception), exception, traceback)
>    1654                     if errors == "skip":
>    1655                         bad_keys.add(key)
> 
> ~/miniconda3/envs/churn/lib/python3.6/site-packages/six.py in reraise(tp, value, tb)
>     690                 value = tp()
>     691             if value.__traceback__ is not tb:
> --> 692                 raise value.with_traceback(tb)
>     693             raise value
>     694         finally:
> 
> /home/saif/miniconda3/envs/churn/lib/python3.6/site-packages/dask/dataframe/utils.py in check_meta()
> 
> ValueError: Metadata mismatch found in `from_delayed`.
> 
> Expected partition of type `DataFrame` but got `DataFrame`

This was run on Ubuntu 16 with Python 3.6 using dask==1.2.2 distributed==1.28.0

[jrbourbeau]

#4819 improves the error message that is raised in the originally posted example

[TomAugspurger]

Closing. I don’t immediately see an easy way to error / warn on this behavior earlier. And hopefully the improved error message would lead the user to see that passing dd.read_csv there is inappropriate.