Mixing
MultiIndex pivot in pandas, long-to-wide problem
I have what I am sure is a simple long to wide problem, but I cannot get the syntax right on how to handle this when I have multiple columns at play.
My data looks like this:
>>> df.head()
s_name geo zip date value
0 A002X zip 00601 2010 1
1 A002Y zip 00601 2010 2
2 A003X zip 00601 2010 3
3 A003Y zip 00601 2010 4
...
. A002X zip 00602 2010 8
. A002Y zip 00602 2010 9
. A003X zip 00602 2010 10
. A003Y zip 00602 2010 11
I would like it to look like this, that is, the contents of the s_name column become new columns, and their values are the 'value' column, grouped by (geo,zip,date):
>>> df2.head()
A002X A002Y A003X A003Y geo zip date
0 1 2 3 4 zip 00601 2010
1 8 9 10 11 zip 00602 2010
2 19 20 21 22 zip 00603 2010
3 31 32 33 45 zip 00604 2010
I've tried a couple like this
>>> df.pivot(columns='s_name', values=["geo","zip","date","value"])
>>> df.pivot(columns='s_name', values=["geo","zip","date","value"]).unstack().reset_index()
I feel like I'm really close, but nothing seems to match up. This, for example, drops the extra non-s_name values:
>>> df.pivot(columns='name', values=["value"]).head()
How do I get the format I'm looking for?
pandas pandas-groupby
asked Jan 2 at 23:17
MittenchopsMittenchops
6,5712268141
add a comment |
I have what I am sure is a simple long to wide problem, but I cannot get the syntax right on how to handle this when I have multiple columns at play.
My data looks like this:
>>> df.head()
s_name geo zip date value
0 A002X zip 00601 2010 1
1 A002Y zip 00601 2010 2
2 A003X zip 00601 2010 3
3 A003Y zip 00601 2010 4
...
. A002X zip 00602 2010 8
. A002Y zip 00602 2010 9
. A003X zip 00602 2010 10
. A003Y zip 00602 2010 11
I would like it to look like this, that is, the contents of the s_name column become new columns, and their values are the 'value' column, grouped by (geo,zip,date):
>>> df2.head()
A002X A002Y A003X A003Y geo zip date
0 1 2 3 4 zip 00601 2010
1 8 9 10 11 zip 00602 2010
2 19 20 21 22 zip 00603 2010
3 31 32 33 45 zip 00604 2010
I've tried a couple like this
>>> df.pivot(columns='s_name', values=["geo","zip","date","value"])
>>> df.pivot(columns='s_name', values=["geo","zip","date","value"]).unstack().reset_index()
I feel like I'm really close, but nothing seems to match up. This, for example, drops the extra non-s_name values:
>>> df.pivot(columns='name', values=["value"]).head()
How do I get the format I'm looking for?
pandas pandas-groupby
asked Jan 2 at 23:17
MittenchopsMittenchops
6,5712268141
1
Does each group {'A002X', ....} have the same number of observations?
– Daniel Mesejo
Jan 3 at 0:13
It ought to. (But data is data....)
– Mittenchops
Jan 3 at 0:18
add a comment |
I have what I am sure is a simple long to wide problem, but I cannot get the syntax right on how to handle this when I have multiple columns at play.
My data looks like this:
>>> df.head()
s_name geo zip date value
0 A002X zip 00601 2010 1
1 A002Y zip 00601 2010 2
2 A003X zip 00601 2010 3
3 A003Y zip 00601 2010 4
...
. A002X zip 00602 2010 8
. A002Y zip 00602 2010 9
. A003X zip 00602 2010 10
. A003Y zip 00602 2010 11
I would like it to look like this, that is, the contents of the s_name column become new columns, and their values are the 'value' column, grouped by (geo,zip,date):
>>> df2.head()
A002X A002Y A003X A003Y geo zip date
0 1 2 3 4 zip 00601 2010
1 8 9 10 11 zip 00602 2010
2 19 20 21 22 zip 00603 2010
3 31 32 33 45 zip 00604 2010
I've tried a couple like this
>>> df.pivot(columns='s_name', values=["geo","zip","date","value"])
>>> df.pivot(columns='s_name', values=["geo","zip","date","value"]).unstack().reset_index()
I feel like I'm really close, but nothing seems to match up. This, for example, drops the extra non-s_name values:
>>> df.pivot(columns='name', values=["value"]).head()
How do I get the format I'm looking for?
pandas pandas-groupby
asked Jan 2 at 23:17
MittenchopsMittenchops
6,5712268141
I have what I am sure is a simple long to wide problem, but I cannot get the syntax right on how to handle this when I have multiple columns at play.
My data looks like this:
>>> df.head()
s_name geo zip date value
0 A002X zip 00601 2010 1
1 A002Y zip 00601 2010 2
2 A003X zip 00601 2010 3
3 A003Y zip 00601 2010 4
...
. A002X zip 00602 2010 8
. A002Y zip 00602 2010 9
. A003X zip 00602 2010 10
. A003Y zip 00602 2010 11
I would like it to look like this, that is, the contents of the s_name column become new columns, and their values are the 'value' column, grouped by (geo,zip,date):
>>> df2.head()
A002X A002Y A003X A003Y geo zip date
0 1 2 3 4 zip 00601 2010
1 8 9 10 11 zip 00602 2010
2 19 20 21 22 zip 00603 2010
3 31 32 33 45 zip 00604 2010
I've tried a couple like this
>>> df.pivot(columns='s_name', values=["geo","zip","date","value"])
>>> df.pivot(columns='s_name', values=["geo","zip","date","value"]).unstack().reset_index()
I feel like I'm really close, but nothing seems to match up. This, for example, drops the extra non-s_name values:
>>> df.pivot(columns='name', values=["value"]).head()
How do I get the format I'm looking for?
pandas pandas-groupby
pandas pandas-groupby
asked Jan 2 at 23:17
MittenchopsMittenchops
6,5712268141
asked Jan 2 at 23:17
MittenchopsMittenchops
6,5712268141
edited Jan 2 at 23:50
Mittenchops
asked Jan 2 at 23:17
MittenchopsMittenchops
6,5712268141
asked Jan 2 at 23:17
MittenchopsMittenchops
6,5712268141
asked Jan 2 at 23:17
MittenchopsMittenchops
6,5712268141
6,5712268141
1
Does each group {'A002X', ....} have the same number of observations?
– Daniel Mesejo
Jan 3 at 0:13
It ought to. (But data is data....)
– Mittenchops
Jan 3 at 0:18
add a comment |
1
Does each group {'A002X', ....} have the same number of observations?
– Daniel Mesejo
Jan 3 at 0:13
It ought to. (But data is data....)
– Mittenchops
Jan 3 at 0:18
1
1
Does each group {'A002X', ....} have the same number of observations?
– Daniel Mesejo
Jan 3 at 0:13
Does each group {'A002X', ....} have the same number of observations?
– Daniel Mesejo
Jan 3 at 0:13
It ought to. (But data is data....)
– Mittenchops
Jan 3 at 0:18
It ought to. (But data is data....)
– Mittenchops
Jan 3 at 0:18
add a comment |
2 Answers
2
active
oldest
votes
What I will do pivot_table
s=df.pivot_table(index=['geo','zip','date'],columns='s_name',values='value',aggfunc='sum')
s
s_name A002X A002Y A003X A003Y
geo zip date
zip 601 2010 1 2 3 4
602 2010 8 9 10 11
#s=s.reset_index()
answered Jan 3 at 0:46
Wen-BenWen-Ben
122k83671
add a comment |
It is necessary to set the dimensions as indices, and unstack the dimension that is desired as columns
This leaves a multi-index in the final output
Example:
# with df as the following
s_name geo zip date value
0 A002X zip 601 2010 1
1 A002Y zip 601 2010 2
2 A003X zip 601 2010 3
3 A003Y zip 601 2010 4
ids = df.columns[:-1].tolist()
df2 = df.set_index(ids).unstack(0)
df2
# outputs:
value
s_name A002X A002Y A003X A003Y
geo zip date
zip 601 2010 1 2 3 4
Then flatten the multiindex, keeping the 2nd level:
df2.columns = df2.columns.get_level_values(1)
answered Jan 3 at 0:18
Haleemur AliHaleemur Ali
12.7k21741
The flatten gives the error:IndexError: Too many levels: Index has only 1 level, not 2df or df2?
– Mittenchops
Jan 3 at 0:24
2
@Mittenchops I believe is df2.
– Daniel Mesejo
Jan 3 at 0:36
@Mittenchops, that was a typo in the last line as pointed out by Daniel Mesejo. The typo has been corrected.
– Haleemur Ali
Jan 3 at 0:47
add a comment |
Your Answer
StackExchange.ifUsing("editor", function () {
StackExchange.using("externalEditor", function () {
StackExchange.using("snippets", function () {
StackExchange.snippets.init();
});
});
}, "code-snippets");
StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "1"
};
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function() {
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled) {
StackExchange.using("snippets", function() {
createEditor();
});
}
else {
createEditor();
}
});
function createEditor() {
StackExchange.prepareEditor({
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader: {
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
},
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
});
}
});
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f54014430%2fmultiindex-pivot-in-pandas-long-to-wide-problem%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
2 Answers
2
active
oldest
votes
2 Answers
2
active
oldest
votes
active
oldest
votes
active
oldest
votes
What I will do pivot_table
s=df.pivot_table(index=['geo','zip','date'],columns='s_name',values='value',aggfunc='sum')
s
s_name A002X A002Y A003X A003Y
geo zip date
zip 601 2010 1 2 3 4
602 2010 8 9 10 11
#s=s.reset_index()
answered Jan 3 at 0:46
Wen-BenWen-Ben
122k83671
add a comment |
What I will do pivot_table
s=df.pivot_table(index=['geo','zip','date'],columns='s_name',values='value',aggfunc='sum')
s
s_name A002X A002Y A003X A003Y
geo zip date
zip 601 2010 1 2 3 4
602 2010 8 9 10 11
#s=s.reset_index()
answered Jan 3 at 0:46
Wen-BenWen-Ben
122k83671
add a comment |
What I will do pivot_table
s=df.pivot_table(index=['geo','zip','date'],columns='s_name',values='value',aggfunc='sum')
s
s_name A002X A002Y A003X A003Y
geo zip date
zip 601 2010 1 2 3 4
602 2010 8 9 10 11
#s=s.reset_index()
answered Jan 3 at 0:46
Wen-BenWen-Ben
122k83671
What I will do pivot_table
s=df.pivot_table(index=['geo','zip','date'],columns='s_name',values='value',aggfunc='sum')
s
s_name A002X A002Y A003X A003Y
geo zip date
zip 601 2010 1 2 3 4
602 2010 8 9 10 11
#s=s.reset_index()
answered Jan 3 at 0:46
Wen-BenWen-Ben
122k83671
answered Jan 3 at 0:46
Wen-BenWen-Ben
122k83671
answered Jan 3 at 0:46
Wen-BenWen-Ben
122k83671
answered Jan 3 at 0:46
Wen-BenWen-Ben
122k83671
122k83671
add a comment |
add a comment |
It is necessary to set the dimensions as indices, and unstack the dimension that is desired as columns
This leaves a multi-index in the final output
Example:
# with df as the following
s_name geo zip date value
0 A002X zip 601 2010 1
1 A002Y zip 601 2010 2
2 A003X zip 601 2010 3
3 A003Y zip 601 2010 4
ids = df.columns[:-1].tolist()
df2 = df.set_index(ids).unstack(0)
df2
# outputs:
value
s_name A002X A002Y A003X A003Y
geo zip date
zip 601 2010 1 2 3 4
Then flatten the multiindex, keeping the 2nd level:
df2.columns = df2.columns.get_level_values(1)
answered Jan 3 at 0:18
Haleemur AliHaleemur Ali
12.7k21741
The flatten gives the error:IndexError: Too many levels: Index has only 1 level, not 2df or df2?
– Mittenchops
Jan 3 at 0:24
2
@Mittenchops I believe is df2.
– Daniel Mesejo
Jan 3 at 0:36
@Mittenchops, that was a typo in the last line as pointed out by Daniel Mesejo. The typo has been corrected.
– Haleemur Ali
Jan 3 at 0:47
add a comment |
It is necessary to set the dimensions as indices, and unstack the dimension that is desired as columns
This leaves a multi-index in the final output
Example:
# with df as the following
s_name geo zip date value
0 A002X zip 601 2010 1
1 A002Y zip 601 2010 2
2 A003X zip 601 2010 3
3 A003Y zip 601 2010 4
ids = df.columns[:-1].tolist()
df2 = df.set_index(ids).unstack(0)
df2
# outputs:
value
s_name A002X A002Y A003X A003Y
geo zip date
zip 601 2010 1 2 3 4
Then flatten the multiindex, keeping the 2nd level:
df2.columns = df2.columns.get_level_values(1)
answered Jan 3 at 0:18
Haleemur AliHaleemur Ali
12.7k21741
The flatten gives the error:IndexError: Too many levels: Index has only 1 level, not 2df or df2?
– Mittenchops
Jan 3 at 0:24
2
@Mittenchops I believe is df2.
– Daniel Mesejo
Jan 3 at 0:36
@Mittenchops, that was a typo in the last line as pointed out by Daniel Mesejo. The typo has been corrected.
– Haleemur Ali
Jan 3 at 0:47
add a comment |
It is necessary to set the dimensions as indices, and unstack the dimension that is desired as columns
This leaves a multi-index in the final output
Example:
# with df as the following
s_name geo zip date value
0 A002X zip 601 2010 1
1 A002Y zip 601 2010 2
2 A003X zip 601 2010 3
3 A003Y zip 601 2010 4
ids = df.columns[:-1].tolist()
df2 = df.set_index(ids).unstack(0)
df2
# outputs:
value
s_name A002X A002Y A003X A003Y
geo zip date
zip 601 2010 1 2 3 4
Then flatten the multiindex, keeping the 2nd level:
df2.columns = df2.columns.get_level_values(1)
answered Jan 3 at 0:18
Haleemur AliHaleemur Ali
12.7k21741
It is necessary to set the dimensions as indices, and unstack the dimension that is desired as columns
This leaves a multi-index in the final output
Example:
# with df as the following
s_name geo zip date value
0 A002X zip 601 2010 1
1 A002Y zip 601 2010 2
2 A003X zip 601 2010 3
3 A003Y zip 601 2010 4
ids = df.columns[:-1].tolist()
df2 = df.set_index(ids).unstack(0)
df2
# outputs:
value
s_name A002X A002Y A003X A003Y
geo zip date
zip 601 2010 1 2 3 4
Then flatten the multiindex, keeping the 2nd level:
df2.columns = df2.columns.get_level_values(1)
answered Jan 3 at 0:18
Haleemur AliHaleemur Ali
12.7k21741
edited Jan 3 at 0:46
answered Jan 3 at 0:18
Haleemur AliHaleemur Ali
12.7k21741
answered Jan 3 at 0:18
Haleemur AliHaleemur Ali
12.7k21741
answered Jan 3 at 0:18
Haleemur AliHaleemur Ali
12.7k21741
12.7k21741
The flatten gives the error:IndexError: Too many levels: Index has only 1 level, not 2df or df2?
– Mittenchops
Jan 3 at 0:24
2
@Mittenchops I believe is df2.
– Daniel Mesejo
Jan 3 at 0:36
@Mittenchops, that was a typo in the last line as pointed out by Daniel Mesejo. The typo has been corrected.
– Haleemur Ali
Jan 3 at 0:47
add a comment |
The flatten gives the error:IndexError: Too many levels: Index has only 1 level, not 2df or df2?
– Mittenchops
Jan 3 at 0:24
2
@Mittenchops I believe is df2.
– Daniel Mesejo
Jan 3 at 0:36
@Mittenchops, that was a typo in the last line as pointed out by Daniel Mesejo. The typo has been corrected.
– Haleemur Ali
Jan 3 at 0:47
The flatten gives the error:
IndexError: Too many levels: Index has only 1 level, not 2 df or df2?– Mittenchops
Jan 3 at 0:24
The flatten gives the error:
IndexError: Too many levels: Index has only 1 level, not 2 df or df2?– Mittenchops
Jan 3 at 0:24
2
2
@Mittenchops I believe is df2.
– Daniel Mesejo
Jan 3 at 0:36
@Mittenchops I believe is df2.
– Daniel Mesejo
Jan 3 at 0:36
@Mittenchops, that was a typo in the last line as pointed out by Daniel Mesejo. The typo has been corrected.
– Haleemur Ali
Jan 3 at 0:47
@Mittenchops, that was a typo in the last line as pointed out by Daniel Mesejo. The typo has been corrected.
– Haleemur Ali
Jan 3 at 0:47
add a comment |
Thanks for contributing an answer to Stack Overflow!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f54014430%2fmultiindex-pivot-in-pandas-long-to-wide-problem%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
1
Does each group {'A002X', ....} have the same number of observations?
– Daniel Mesejo
Jan 3 at 0:13
It ought to. (But data is data....)
– Mittenchops
Jan 3 at 0:18