Thanks Wes, having a flag would be just fine. On Tue, Dec 26, 2017 at 2:25 PM, Wes McKinney <wesmck...@gmail.com> wrote:
> hi Brian, > > It does not, but we could add a flag for this: > > https://issues.apache.org/jira/browse/ARROW-1949 > > - Wes > > On Fri, Dec 22, 2017 at 6:56 PM, Bryan Cutler <cutl...@gmail.com> wrote: > > Hi devs, > > > > I am creating a pyarrow.Array from a pandas.Series and when specifying > the > > data type, if the value will be truncated an error is raised. For > example: > > > > import pyarrow as pa > > import pandas as pd > > s = pd.Series([pd.Timestamp(1)]) > > arr = pa.Array.from_pandas(s, type=pa.timestamp('us')) > > > > /home/bryan/miniconda2/lib/python2.7/site-packages/pyarrow/error.pxi in > > pyarrow.lib.check_status > > (/arrow/python/build/temp.linux-x86_64-2.7/lib.cxx:8270)() > > 75 message = frombytes(status.message()) > > 76 if status.IsInvalid(): > > ---> 77 raise ArrowInvalid(message) > > 78 elif status.IsIOError(): > > 79 raise ArrowIOError(message) > > > > ArrowInvalid: Casting from timestamp[ns] to timestamp[us] would lose > data: 1 > > > > > > If I do the casting after creating the pyarrow.Array like > > > > pa.Array.from_pandas(s).cast(pa.timestamp('us'), safe=False) > > > > then it is fine. My question is does pyarrow.Array.from_pandas only use > > safe casting or is there a way to specify that truncation is ok? > > > > Thanks, > > Bryan >
