[jira] [Updated] (ARROW-1354) Pyarrow Segfault with Mixed-Type Categories

["Nick White (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/ARROW-1354?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Nick White updated ARROW-1354:
------------------------------
    Description: 
I'm not sure what the expected behaviour should be, but it's not segfaulting!

{code}
import pandas as pd
import pyarrow as pa
print("Pandas: %s" % pd.__version__)
print("Pyarrow: %s" % pa.__version__)
df = pd.DataFrame({"B": (1, 'A')})
df = df.assign(B=df.B.astype("category"))
pa.Table.from_pandas(df)  # this segfaults!
{code}

Versions:

{code}
Pandas: 0.20.1
Pyarrow: 0.5.0
{code}


  was:
I'm not sure what the expected behaviour should be, but it's not segfaulting!

```
import pandas as pd
import pyarrow as pa
print("Pandas: %s" % pd.__version__)
print("Pyarrow: %s" % pa.__version__)
df = pd.DataFrame({"B": (1, 'A')})
df = df.assign(B=df.B.astype("category"))
pa.Table.from_pandas(df)  # this segfaults!
```

Versions:

```
Pandas: 0.20.1
Pyarrow: 0.5.0
```



> Pyarrow Segfault with Mixed-Type Categories
> -------------------------------------------
>
>                 Key: ARROW-1354
>                 URL: https://issues.apache.org/jira/browse/ARROW-1354
>             Project: Apache Arrow
>          Issue Type: Bug
>          Components: Python
>    Affects Versions: 0.5.0
>            Reporter: Nick White
>
> I'm not sure what the expected behaviour should be, but it's not segfaulting!
> {code}
> import pandas as pd
> import pyarrow as pa
> print("Pandas: %s" % pd.__version__)
> print("Pyarrow: %s" % pa.__version__)
> df = pd.DataFrame({"B": (1, 'A')})
> df = df.assign(B=df.B.astype("category"))
> pa.Table.from_pandas(df)  # this segfaults!
> {code}
> Versions:
> {code}
> Pandas: 0.20.1
> Pyarrow: 0.5.0
> {code}



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)