currently stuck with something I hope to find an answer for in this forum:
I have a df with multiple columns containing URLs. My index column are URLs as well.
AIM: I'd like to replace df values across all columns with np.NaN if the number of "/" (count()) in the index is not equal to the number of "/" (count()) in the values of each individual of of the other columns
E.x.
First, you need one column to compare to.
counts = df['id_url'].str.count('/')
Then you evaluate all the rows at once.
mask = df.str.count('/') == counts
Then we want to to show rows where all the values are equal.
mask = mask.all(axis=1)
Now we have a mask for where every value is equal, we can use the not operator to filter for those where at least one column is not equal.
df.loc[~mask, :] = np.nan # replaces every value in the row with np.nan
Related
I have two Dataframes, one containing my data read in from a CSV file and another that has the data grouped by all of the columns but the last and reindexed to contain a column for the count of the size of the groups.
df_k1 = pd.read_csv(filename, sep=';')
columns_for_groups = list(df_k1.columns)[:-1]
k1_grouped = df_k1.groupby(columns_for_groups).size().reset_index(name="Count")
I need to create a series such that every row(i) in the series corresponds to row(i) in my original Dataframe but the contents of the series need to be the size of the group that the row belongs to in the grouped Dataframe. I currently have this, and it works for my purposes, but I was wondering if anyone knew of a faster or more elegant solution.
size_by_row = []
for row in df_k1.itertuples():
for group in k1_grouped.itertuples():
if row[1:-1] == group[1:-1]:
size_by_row.append(group[-1])
break
group_size = pd.Series(size_by_row)
I have two dataframes.
df1 has an index list made of strings like (row1,row2,..,rown) and a column list made of strings like (col1,col2,..,colm) while df2 has k rows and 3 columns (char_1,char_2,value). char_1 contains strings like df1 indexes while char_2 contains strings like df1 columns. I only want to assign the df2 value to df1 in the right position. For example if the first row of df2 reads ['row3','col1','value2'] I want to assign value2 to df1 in the position ([2,0]) (third row and first column).
I tried to use two functions to slide rows and columns of df1:
def func1(val):
# first I convert the series to dataframe
val=val.to_frame()
val=val.reset_index()
val=val.set_index('index') # I set the index so that it's the right column
def func2(val2):
try: # maybe the combination doesn't exist
idx1=list(cou.index[df2[char_2]==(val2.name)]) #val2.name reads col name of df1
idx2=list(cou.index[df2[char_1]==val2.index.values[0]]) #val2.index.values[0] reads index name of df1
idx= list(reduce(set.intersection, map(set, [idx1,idx2])))
idx=int(idx[0]) # final index of df2 where I need to take value to assign to df1
check=1
except:
check=0
if check==1: # if index exists
val2[0]=df2['value'][idx] # assign value to df1
return val2
val=val.apply(func2,axis=1) #apply the function for columns
val=val.squeeze() #convert again to series
return val
df1=df1.apply(func1,axis=1) #apply the function for rows
I made the conversion inside func1 because without this step I wasn't able to work with series keeping index and column names so I wasn't able to find the index idx in func2.
Well the problem is that it takes forever. df1 size is (3'600 X 20'000) and df2 is ( 500 X 3 ) so it's not too much. I really don't understand the problem.. I run the code for the first row and column to check the result and it's fine and it takes 1 second, but now for the entire process I've been waiting for hours and it's still not finished.
Is there a way to optimize it? As I wrote in the title I only need to run a function that keeps column and index names and works sliding the entire dataframe. Thanks in advance!
I have a column in a dataframe where all the values are lists (list of one item usually for each row). So, I would like to use get_dummies to one hot encode all the values. However, there may be a few rows where there is not a value for the column. I have seen it originally as a nan and then I have replaced that nan with an empty list, but in either case I do not see 0 and 1s for the result for the get_dummies, but rather each generated column is blank (I would expect each generated column to be 0).
How do I get get_dummies to work with an empty list?
# create column from dict where value will be a list
X['sponsor_list'] = X['bill_id'].map(sponsor_non_plaw_dict)
# line to replace nan in sponsor_list column with empty list
X.loc[X['sponsor_list'].isnull(),['sponsor_list']] = X.loc[X['sponsor_list'].isnull(),'sponsor_list'].apply(lambda x: [])
# use of get_dummies to encode the sponsor_list column
X = pd.concat([X, pd.get_dummies(X.sponsor_list.apply(pd.Series).stack()).sum(level=0)], axis=1)
Example:
111th-congress_senate-bill_3695.txt False ['Menendez,_Robert_[D-NJ].txt']
112th-congress_house-bill_3630.txt False []
111th-congress_senate-bill_852.txt False ['Vitter,_David_[R-LA].txt']
114th-congress_senate-bill_2832.txt False
['Isakson,_Johnny_[R-GA].txt']
107th-congress_senate-bill_535.txt False ['Bingaman,_Jeff_[D-NM].txt']
I want to one hot encode on the third column. That particular data item in the 2nd row has no person associated it with them, so I need that row to be encoded with all 0s. The reason I need the third column to be a list is because I need to do this to a related column as well where I need to have [0,n] values where n can be 5 or 10 or even 20.
X['sponsor_list'] = X['bill_id'].map(sponsor_non_plaw_dict)
X.loc[X['sponsor_list'].isnull(),['sponsor_list']] = X.loc[X['sponsor_list'].isnull(),'sponsor_list'].apply(lambda x: [])
mlb = MultiLabelBinarizer()
X = X.join(pd.DataFrame(mlb.fit_transform(X.pop('sponsor_list')),
columns=mlb.classes_,
index=X.index))
I used a MultiLabelBinarizer to capture what I was trying to do. I still replace nan with empty list before applying, but then I fit_transform to create the 0/1 values which can result in no 1's in a row, or many 1's in a row.
I have a dataframe at hourly level with several columns. I want to extract the entire rows (containing all columns) of the 10 top values of a specific column for every year in my dataframe.
so far I ran the following code:
df = df.groupby([df.index.year])['totaldemand'].apply(lambda grp: grp.nlargest(10)))
The problem here is that I only get the top 10 values for each year of that specific column and I lose the other columns. How can I do this operation and having the corresponding values of the other columns that correspond to the top 10 values per year of my 'totaldemand' column?
We usually do head after sort_values
df = df.sort_values('totaldemand',ascending = False).groupby([df.index.year])['totaldemand'].head(10)
nlargest can be applied to each group, passing the column to look for
largest values.
So run:
df.groupby([df.index.year]).apply(lambda grp: grp.nlargest(3, 'totaldemand'))
Of course, in the final version replace 3 with your actual value.
Get the index of your query and use it as a mask on your original df:
idx = df.groupby([df.index.year])['totaldemand'].apply(lambda grp: grp.nlargest(10))).index.to_list()
df.iloc[idx,]
(or something to that extend, I can't test now without any test data)
I am selecting row by row as follows:
for i in range(num_rows):
row = df.iloc[i]
as a result I am getting a Series object where row.index.values contains names of df columns.
But I wanted instead dataframe with only one row having dataframe columns in place.
When I do row.to_frame() instead of 1x85 dataframe (1 row, 85 cols) I get 85x1 dataframe where index contains names of columns and row.columns
outputs
Int64Index([0], dtype='int64').
But all I want is just original data-frame columns with only one row. How do I do it?
Or how do I convert row.index values to row.column values and change 85x1 dimension to 1x85
You just need to adding T
row.to_frame().T
Also change your for loop with adding []
for i in range(num_rows):
row = df.iloc[[i]]