-
-
Notifications
You must be signed in to change notification settings - Fork 18.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
BUG: concat of Series w/o names #10698 #10723
Conversation
columns = [] | ||
for i in range(len(data)): | ||
columns.append(Series(data[i]).name if Series(data[i]).name is not None else i) | ||
tmpdf.columns = columns |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
use a counter here - first missing should be 0
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hi jreback. I see your point. I've erroneously believed we wanted to use the column index as the column name in case the parent Series didn't have a valid name. I'm going to update the pull request to include the change you suggested.
Apologies in advance if my code is not always the neatest. I'm new to software development, although I use pandas everyday for data analysis. I thought actively participating to the project could be a good way for me to learn how to code properly and become a better programmer.
here are some helpful hints FYI: http://pandas.pydata.org/pandas-docs/stable/contributing.html |
Apologies I pushed a few "fixes" for some unit tests which were failing. Looking at the reason why those tests were failing I've noticed it was because of the use of the argument 'keys' in the function 'concat'. Most people are using it to pass the column names, which is something I personally don't do and therefore I didn't consider when writing the code. I'm going to reset my local branch to commit 'b9cba86ec1c11c200c44c62d0a300b7f326a12e4' and make the necessary changes to allow this case. Before sending the next pull request I'm going to make sure all nosetests pass correctly. This should also fix the build issue. Apologies again, this is due to lack of experience. I'll make sure to learn from my mistake. |
@@ -1797,6 +1797,15 @@ def test_concat_dataframe_keys_bug(self): | |||
self.assertEqual(list(result.columns), [('t1', 'value'), | |||
('t2', 'value')]) | |||
|
|||
def test_concat_series_partial_columns_names(self): |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
add the issue number as a comment here
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've added the reference to the GitHub issue
@jreback Thanks for spending the time to review my pull request. I'm going to start working on another bug now. |
pls add a whatsnew note. This is just a bug fix, but I think should have a mini-example. pls have squash as well. |
I think I made a mistake. I rebased my local version because was missing some of the recent updates in the upstream version and (after solving some merging conflicts) I pushed. I see there is an alert message saying "This branch has conflicts that must be resolved" on my pull request in GitHub. Apologies for that. Is there a way I can solve this? |
you need to rebase / squash. If you have a conflict it will show up when you do this. Conflicts are normal and happen because others have changed code that you are changing. See the contributing docs here |
Okay! So it should be fine right? |
Conflicts are fine, as in they're not a sign you've done anything wrong. But they do need to be resolved before merge - the docs describe how to squash & merge. Cheers |
Okay guys, I think I did it. Please let me know if I did something wrong. Apologies again for my lack of experience. |
@@ -152,6 +153,30 @@ Other enhancements | |||
s.drop_duplicates(keep=False) | |||
|
|||
|
|||
- ``concat`` will now inherit the existing series names (even when some are missing), if new ones are not provided through the ``keys`` argument (:issue:`10698`). |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I would say something like: will use existing Series
names if provided.
might need a doc-example http://pandas.pydata.org/pandas-docs/stable/merging.html#more-concatenating-with-group-keys to explain how |
can you update and add a doc example? |
Sure! I think I messed up with my branch because lots of test are failing after I rebased last day. I'll fix this ASAP. |
I see the branch has conflicts. I'm assuming in this sort of situations I need to fix those myself right? In the effort of solving them using
Then I update the local branch:
The In the picture above the document on the left is my local, the one in the middle is the result of the merge, and the one on the right is the remote. In this occasions I tend to pick from the remote everything I don't have in my local branch, because I assume these are coming from other submitted pull requests or code merged while I was working on my local branch. What about the code which is in my local, hasn't been written by me, and is not in the remote? Should I discard it? |
|
||
.. ipython:: python | ||
|
||
foo = pd.Series([1,2], name='foo') |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
just create all of these in a separate python block above previous behavior. Then no need to show them twice. (obviously show the results in the previous behavior section as a code-block then in an ipython block in new behavior).
need to rebase / squash. generally you won't have conflicts. other people add code and should be straightforward to accept. |
|
||
.. ipython:: python | ||
|
||
s3 = Series([0, 1, 2, 3], name='foo') |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Can you use pd.
before all Series and concat calls?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We are converting the docs (already partly, but maybe this one not)
Seems you have a bit too much changes in the whatsnew file? |
Joris, my bad, I messed up when merging. I think I've solved all the issues now. |
|
||
import pandas.core.common as com | ||
|
||
import pandas.lib as lib | ||
import pandas.algos as algos |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
extra import?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@jreback You're right. I've fixed!
can you update according to comments |
BUG: concat of Series w/o names #10698
@IamGianluca awesome job! thanks! |
closes #10698
Let the result of 'concat' to inherit the parent Series' names. The Series' name (if present) will be used as the resulting DataFrame column name. When only one of the Series has a valid name, the resulting DataFrame will inherit the name only, and use a column name for the other columns the column index value.