Fix CI failures found while testing SDK with airflow 2.5.0 #1337
Assignees
Labels
Milestone
Comments
kaxil
added a commit
that referenced
this issue
Dec 1, 2022
This test was checking a wrong thing, see below:
```
In [1]: import pandas
In [2]: df = pandas.DataFrame({"Numbers": [1, 2, 3], "Colors": ["red", "white", "blue"]})
In [3]: df.columns
Out[3]: Index(['Numbers', 'Colors'], dtype='object')
In [4]: cols = list(df.columns)
In [5]: cols.sort()
In [6]: df.columns == ["Colors", "Numbers"]
Out[6]: array([False, False])
```
It was returning a numpy array with `False, False` values, instead the test actually was meant to test whether the column case was preserved or not.
part of #1337
kaxil
added a commit
that referenced
this issue
Dec 1, 2022
This test was checking a wrong thing, see below:
```python
In [1]: import pandas
In [2]: df = pandas.DataFrame({"Numbers": [1, 2, 3], "Colors": ["red", "white", "blue"]})
In [3]: df.columns
Out[3]: Index(['Numbers', 'Colors'], dtype='object')
In [4]: cols = list(df.columns)
In [5]: cols.sort()
In [6]: df.columns == ["Colors", "Numbers"]
Out[6]: array([False, False])
```
It was returning a numpy array with `False, False` values, instead the
test actually was meant to test whether the column case was preserved or
not.
part of #1337
### Checklist
- [x] Created tests which fail without the change (if possible)
- [x] Extended the README / documentation, if necessary
2 tasks
feluelle
changed the title
feluelle
pushed a commit
that referenced
this issue
Dec 5, 2022
Currently, the `LoadFileOperator` converts dataframe to a File object when returning. However, this should be handled by the Custom XCom backend and not in `LoadFileOperator`. Otherwise, we won't be consistent with dataframe and transform decorators who, too returns dataframes. From Airflow 2.5 and above, this should be handled automatically, for older versions, users should use our Custom XCom backend so that their XCom table isn't fully of GBs of data. part of #1337 ## Does this introduce a breaking change? No, if you are using our XCom backend. Yes, if you are not! but this fixes the behaviour and makes it consistent with other decorators and operators that return dataframes ### Checklist - [x] Created tests which fail without the change (if possible) - [x] Extended the README / documentation, if necessary
|
Fixed by #1317 |
kaxil
added a commit
that referenced
this issue
Dec 13, 2022
This test was checking a wrong thing, see below:
```python
In [1]: import pandas
In [2]: df = pandas.DataFrame({"Numbers": [1, 2, 3], "Colors": ["red", "white", "blue"]})
In [3]: df.columns
Out[3]: Index(['Numbers', 'Colors'], dtype='object')
In [4]: cols = list(df.columns)
In [5]: cols.sort()
In [6]: df.columns == ["Colors", "Numbers"]
Out[6]: array([False, False])
```
It was returning a numpy array with `False, False` values, instead the
test actually was meant to test whether the column case was preserved or
not.
part of #1337
### Checklist
- [x] Created tests which fail without the change (if possible)
- [x] Extended the README / documentation, if necessary
(cherry picked from commit 97de941)
kaxil
added a commit
that referenced
this issue
Dec 13, 2022
Currently, the `LoadFileOperator` converts dataframe to a File object when returning. However, this should be handled by the Custom XCom backend and not in `LoadFileOperator`. Otherwise, we won't be consistent with dataframe and transform decorators who, too returns dataframes. From Airflow 2.5 and above, this should be handled automatically, for older versions, users should use our Custom XCom backend so that their XCom table isn't fully of GBs of data. part of #1337 ## Does this introduce a breaking change? No, if you are using our XCom backend. Yes, if you are not! but this fixes the behaviour and makes it consistent with other decorators and operators that return dataframes ### Checklist - [x] Created tests which fail without the change (if possible) - [x] Extended the README / documentation, if necessary (cherry picked from commit 1283394)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Describe the bug
Failing jobs on CI:
Related to #1334 and #1331 and #1319
The text was updated successfully, but these errors were encountered: