Arnaudno opened a new issue, #45598:
URL: https://github.com/apache/arrow/issues/45598
### Describe the bug, including details regarding any error messages,
version, and platform.
Hello,
with the following example :
```
import pyarrow as pa
import pyarrow.parquet as pq
import pyarrow.compute as pc
import pandas as pd
df = pd.DataFrame({'date' : pd.date_range(start =
pd.to_datetime('2025-02-01'),end = pd.to_datetime('2025-02-20'))})
df.to_parquet('dates.parquet')
timestamp_filter = pd.Timestamp('2025-02-18')
string_filter = str(timestamp_filter)
```
**'=' works with the Timestamp but not the string :**
```
In:
filtered_df = pq.read_table(
source = 'dates.parquet',
filters = [('date','=',timestamp_filter)]
)
print(filtered_df)
Out:
pyarrow.Table
date: timestamp[ns]
----
date: [[2025-02-18 00:00:00.000000000]]
In:
filtered_df = pq.read_table(
source = 'dates.parquet',
filters = [('date','=',string_filter)]
)
Out:
ArrowNotImplementedError: Function 'equal' has no kernel matching input
types (timestamp[ns], string)
```
**however 'in' works with both the Timestamp and the string :**
```
In:
filtered_df = pq.read_table(
source = 'dates.parquet',
filters = [('date','in',[timestamp_filter])]
)
print(filtered_df)
Out:
pyarrow.Table
date: timestamp[ns]
----
date: [[2025-02-18 00:00:00.000000000]]
In:
filtered_df = pq.read_table(
source = 'dates.parquet',
filters = [('date','in',[string_filter])]
)
print(filtered_df)
Out:
pyarrow.Table
date: timestamp[ns]
----
date: [[2025-02-18 00:00:00.000000000]]
```
```
print(pd.__version__)
print(pa.__version__)
2.2.3
17.0.0
```
My problem is that my filter is stored in JSON format (in a dash.store
component https://dash.plotly.com/dash-core-components/store); which
stringifies pd.Timestamp values.
It would make it easier to manage if the single value operators '=' , '>',
'<' could behave like the list like operators 'in' and 'not in'.
Many thanks,
Rgds
### Component(s)
Python
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
