TypeError: unhashable type

[visayang2005]

when i commit the code below:

clustermodel=kprototypes.KPrototypes(n_clusters=6, init='Cao', verbose=2)
clustermodel.fit(sd1,categorical=[0,1])

it can't run with the " TypeError: unhashable type", but kmode is ok. could you pls tell me how can solve this problem? Thank very much!

[nicodv]

Please post a full traceback, so that I can locate the error.

…

On Apr 16, 2017 11:47 PM, "visayang2005" ***@***.***> wrote: when i commit the code below: clustermodel=kprototypes.KPrototypes(n_clusters=6, init='Cao', verbose=2) clustermodel.fit(sd1,categorical=[0,1]) it can't run with the " TypeError: unhashable type", but kmode is ok. could you pls tell me how can solve this problem? Thank very much! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#40>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADi0g_l8lQRXo4nKs2vGJOYZM6txNvJaks5rwwrugaJpZM4M-67U> .

[visayang2005]

ok, my full code like this,

from kmodes import kmodes
from kmodes import kprototypes
import pandas as pd
from sklearn import cluster

the first two varibles of dataframe sd1 is categorical, other varibles is continus

sd1=sd[['state','newmode_id','at_price_scaled','day_diff_scaled', 'time_avg_rate_scaled', 'total_dealnum_scaled', 'total_dealprice_scaled', 'total_dealnum_fws_scaled']]

clustermodel=kprototypes.KPrototypes(n_clusters=6, init='Cao', verbose=2)
cluser=clustermodel.fit_predict(sd1,categorical=[0,1])

[nicodv]

Can you post the full error too, please?

…

On Apr 16, 2017 11:57 PM, "visayang2005" ***@***.***> wrote: ok, my full code like this, from kmodes import kmodes from kmodes import kprototypes import pandas as pd from sklearn import cluster the first two varibles of dataframe sd1 is categorical, other varibles is continus sd1=sd[['state','newmode_id','at_price_scaled','day_diff_scaled', 'time_avg_rate_scaled', 'total_dealnum_scaled', 'total_dealprice_scaled', 'total_dealnum_fws_scaled']] clustermodel=kprototypes.KPrototypes(n_clusters=6, init='Cao', verbose=2) cluser=clustermodel.fit_predict(sd1,categorical=[0,1]) — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#40 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADi0gzKfclSR07U3GRBHwuArpjW7hneAks5rww1GgaJpZM4M-67U> .

[visayang2005]

sorry, pls ignore before, use this:

cluser=clustermodel.fit_predict(sd1,categorical=[0,1])
Traceback (most recent call last):

File "", line 1, in
cluser=clustermodel.fit_predict(sd1,categorical=[0,1])

File "build\bdist.win-amd64\egg\kmodes\kmodes.py", line 374, in fit_predict
return self.fit(X, **kwargs).labels_

File "build\bdist.win-amd64\egg\kmodes\kprototypes.py", line 415, in fit
self.verbose)

File "build\bdist.win-amd64\egg\kmodes\kprototypes.py", line 154, in k_prototypes
Xnum, Xcat = _split_num_cat(X, categorical)

File "build\bdist.win-amd64\egg\kmodes\kprototypes.py", line 47, in _split_num_cat
if ii not in categorical]]).astype(np.float64)

File "D:\Anaconda2\lib\site-packages\pandas\core\frame.py", line 2059, in getitem
return self._getitem_column(key)

File "D:\Anaconda2\lib\site-packages\pandas\core\frame.py", line 2066, in _getitem_column
return self._get_item_cache(key)

File "D:\Anaconda2\lib\site-packages\pandas\core\generic.py", line 1384, in _get_item_cache
res = cache.get(item)

TypeError: unhashable type

[visayang2005]

Hi Nico,

The full error like this, can you help me to solve it, I really appreciate your help.

cluser=clustermodel.fit_predict(sd1,categorical=[0,1])
Traceback (most recent call last):

File "", line 1, in
cluser=clustermodel.fit_predict(sd1,categorical=[0,1])

File "build\bdist.win-amd64\egg\kmodes\kmodes.py", line 374, in fit_predict
return self.fit(X, **kwargs).labels_

File "build\bdist.win-amd64\egg\kmodes\kprototypes.py", line 415, in fit
self.verbose)

File "build\bdist.win-amd64\egg\kmodes\kprototypes.py", line 154, in k_prototypes
Xnum, Xcat = _split_num_cat(X, categorical)

File "build\bdist.win-amd64\egg\kmodes\kprototypes.py", line 47, in _split_num_cat
if ii not in categorical]]).astype(np.float64)

File "D:\Anaconda2\lib\site-packages\pandas\core\frame.py", line 2059, in getitem
return self._getitem_column(key)

File "D:\Anaconda2\lib\site-packages\pandas\core\frame.py", line 2066, in _getitem_column
return self._get_item_cache(key)

File "D:\Anaconda2\lib\site-packages\pandas\core\generic.py", line 1384, in _get_item_cache
res = cache.get(item)

TypeError: unhashable type

[nicodv]

Try presenting a numpy array to the algorithm, instead of a pandas DataFrame:

cluster=clustermodel.fit_predict(sd1.values,categorical=[0,1])

[nicodv]

Since it seems reasonable to have users present pandas DataFrames as inputs to the algorithm, it's probably a good idea to include a check (e.g., 'pandas' in str(X.__class__)) and do a X.values if True.

Alternatively, make sure all operations (such as the one that caused this bug) support DataFrames, but that's hard to implement and check consistently.

[visayang2005]

Thanks, Nico. My problem is solved by using a numpy array to the algorithm.

[nicodv]

This should no longer occur: a30da80

[cibic89]

This is not fixed in the latest version. Added the .values to the dataframe and worked like a charm.
The type error is "TypeError: unhashable type: slice"

Snippet from kprototypes:

# Convert pandas objects to numpy arrays.
if 'pandas' in str(X.__class__):
X = X.values

[nicodv]

Fixed for real by: #67