[
https://issues.apache.org/jira/browse/ARROW-11183?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17267449#comment-17267449
]
Ivan Smirnov commented on ARROW-11183:
--------------------------------------
[~csun] Here's a Python example that works:
{code:java}
import pandas as pd
import pyarrow.parquet as pq
df = pd.DataFrame(dict(x=[pd.Timestamp.now() for _ in range(10)]))
table = pa.table(df)
pq.write_table(table, 'timestamps.parquet', version='2.0')
assert (pq.read_table('timestamps.parquet').to_pandas() == df).all().all()
{code}
What is the Rust equivalent then?
Note: the table's schema shows up as
{code:java}
pyarrow.Table
x: timestamp[ns]
{code}
> [Rust] [Parquet] LogicalType::TIMESTAMP_NANOS missing
> -----------------------------------------------------
>
> Key: ARROW-11183
> URL: https://issues.apache.org/jira/browse/ARROW-11183
> Project: Apache Arrow
> Issue Type: Improvement
> Reporter: Ivan Smirnov
> Priority: Major
>
> There's UnitTime::NANOS in parquet-format, but no nanosecond timestamp
> support (seemingly) in schema's LogicalType. What is needed to add support
> for nanosecond timestamps in Rust Parquet?
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
