I have a dataframe with millions of rows with unique indexes and a column('b') that has several repeated values.
I would like to generate a dataframe without the duplicated data but I do not want to lose the index information. I want the new dataframe to have an index that is a concatenation of the indexes ("old_index1,old_index2") where 'b' had duplicated values but remains unchanged for rows where 'b' had unique values. The values of the 'b' column should remain unchanged like in a keep=first strategy. Example below.
Input dataframe:
df = pd.DataFrame(data = [[1,"non_duplicated_1"],
[2,"duplicated"],
[2,"duplicated"],
[3,"non_duplicated_2"],
[4,"non_duplicated_3"]],
index=['one','two','three','four','five'],
columns=['a','b'])
desired output:
a b
one 1 non_duplicated_1
two,three 2 duplicated
four 3 non_duplicated_2
five 4 non_duplicated_3
The actual dataframe is quite large so I would like to avoid non-vectorized operations.
I am finding this surprisingly difficult...Any ideas?
You can use transform on the index column (after you use reset_index). Then, drop duplicates in column b:
df.index = df.reset_index().groupby('b')['index'].transform(','.join)
df.drop_duplicates('b',inplace=True)
>>> df
a b
index
one 1 non_duplicated_1
two,three 2 duplicated
four 3 non_duplicated_2
five 4 non_duplicated_3
Setup
dct = {'index': ','.join, 'a': 'first'}
You can reset_index before using groupby, although it's unclear to me why you want this:
df.reset_index().groupby('b', as_index=False, sort=False).agg(dct).set_index('index')
b a
index
one non_duplicated_1 1
two,three duplicated 2
four non_duplicated_2 3
five non_duplicated_3 4
Related
I have the same dataset but over different weeks (so later weeks contain new rows). I want to append the new rows to the original dataframe to create one big dataframe with all unique rows and no duplicates. I can't just take the last week because some get deleted over the weeks.
I tried to use the following code but somehow my final_info dataframe still contains some non-unique values
final_info = data[list(data.keys())[-1]]['all_info']
for week in reversed(data.keys()):
df_diff = pd.concat([data[week]['all_info'],final_info]).drop_duplicates(subset='project_slug',
keep=False)
final_info = final_info.append(df_diff).reset_index(drop=True)
Does somebody see where it goes wrong?
if I understand your question, you are just trying to add the unique rows from one dataframe to another dataframe. I don't think there is any need to iterate through the keys like you are doing. There is an example on this question that I think can help you and i think it is conceptually easier to follow 1. I'll try to walk through an example to make it more clear.
So if you have a dataframe A:
col1 col2
1 2
2 3
3 4
and a dataframe B:
col1 col2
1 2
2 3
6 4
These two dataframes have the same first two rows but have different last rows. If you wanted to get all the unique rows into one dataframe you could first get all the unique rows from just one of the dataframes. So for this example you could get the unique row in dataframe B, lets call it df_diff in this example. The code to do this would be
df_diff = B[~B.col1.isin(A.col1)]
output: col1 col2
6 4
This above line of code makes whats called a boolean mask and then negates using ~ so that you get all rows in dataframe B where the col1 value is not in dataframe A.
You could then merge this dataframe, df_diff, with the first dataframe A. We can call this df_full. This step is done with:
df_full = pd.concat([A, df_diff], ignore_index=True)
The ignore_index=True just resets the index of the resulting dataframe. This will give you:
col1 col2
1 2
2 3
3 4
6 4
Now the above dataframe has the new row in dataframe B plus the original rows from dataframe A.
I think this would work for your situation and may be less lines of code.
I wanted to see how to express counts of unique values of column B for each unique value in column A where corresponding column C >0.
df:
A B C
1 10 0
1 12 3
2 3 1
I tried this but its missing the where clause to filter for C>0. How do I add it?
df.groupby(['A'])['B'].apply(lambda b : b.astype(int).nunique())
Let's first start by creating the dataframe that OP mentions in the question
import pandas as pd
df = pd.DataFrame({'A': [1,1,2], 'B': [10,12,3], 'C': [0,3,1]})
Now, in order to achieve what OP wants, there are various options. One of the ways is by selecting the df where column C is greater than 0, then use pandas.DataFrame.groupby to group by column A. Finally, use pandas.DataFrame.nunique to count the unique values of column B. In one line it would look like the following
count = df[df['C'] > 0].groupby('A')['B'].nunique()
[Out]:
A
1 1
2 1
If one wants to sum every number in of unique items that satisfy the condition, from the dataframe count, one can simply do
count = count.sum()
[Out]:
2
Assuming one wants to do everything in one line, one can use pandas.DataFrame.sum as
count = df[df['C'] > 0].groupby('A')['B'].nunique().sum()
[Out]:
2
I am trying to multiply a few specific columns by a portion of multiple rows and creating a new column from every result. I could not really find an answer to my question in previous stackoverflow questions or on google, so maybe one of you can help.
I would like to point out that I am quite the beginner in Python, so apologies ahead for any obvious questions or strange code.
This is how my DataFrame currently looks like:
So, for the column Rank of Hospital by Doctor_1, I want to multiply all its numbers by the values of the first row of column Rank of Doctor by Hospital_1 until column Rank of Doctor by Hospital_10. Which would result in:
1*1
2*1
3*1
4*4
...
and so on.
I want to do this for every Doctor_ column. So for Doctor_2 its values should be multiplied by the second row of all those ten columns (Rank of Doctor by Hospital_. Doctor_3, multiplied by the third row etc.
So far, I have transposed the Rank of Doctor by Hospital_ columns in a new DataFrame:
and tried to multiply this by a DataFrame of the Rank of Hospital by Doctor_ columns. Here the first column of the first df should be multiplied by the first column of the second df. (and second column * second column, etc.):
But my current formula
preferences_of_doctors_and_hospitals_doctors_ranking.mul(preferences_of_doctors_and_hospitals_hospitals_ranking_transposed)
is obviously not working:
Does anybody know what I am doing wrong and how I could fix this? Maybe I could write a for loop so that a new column is created for every multiplication of columns? So Multiplication_column_1 of DF3 = Column 1 of DF1 * Column 1 of DF2 and Multiplication_column_2 of DF3 = Column 2 of DF1 * Column 2 of DF2.
Thank you in advance!
Jeff
You can multiple 2d arrays created by filtering column with filter and values first:
arr = df.filter(like='Rank of Hospital by').values * df.filter(like='Rank of Doctor by').values
Or:
arr = (preferences_of_doctors_and_hospitals_doctors_ranking.values *
preferences_of_doctors_and_hospitals_hospitals_ranking_transposed.values)
Notice - necessary is same ordering of columns, same length of columns names and index in both filtered DataFrames.
Get 2d array, so create DataFrame by constructor and join to original:
df = df.join(pd.DataFrame(arr, index=df.index).add_prefix('Multiplied '))
df = pd.DataFrame({"A":[1,2,3,4,5], "B":[6,7,8,9,10]})
df["mul"] = df["A"] * df["B"]
print(df)
Output:
A B mul
0 1 6 6
1 2 7 14
2 3 8 24
3 4 9 36
4 5 10 50
If I understood the question correctly I think you way over complicated it.
You can just create another column telling pandas to give it the value of first column multiplied by second column.
More similar to your specific case with more than 2 columns:
df = pd.DataFrame({"A":[1,2,3,4,5], "B":[6,7,8,9,10], "C":[11,12,13,14,15]})
df["mul"] = df["A"] * df["B"] * df["C"]
For example, I have a dataframe called dat, then I want to apply a function on each column of the dataframe, if the return value is Ture, then keep this column and turn to next column, if the return value is False, then drop this column and turn to next column.
I know I can write a for loop to do this, but is there a efficient way to do this?
You could do it like this using boolean index on df.columns:
I want to drop all columns where the 'sum' for simplicity is greater than 50
df = pd.DataFrame({'A':[2,4,6,8],'B':[101,102,102,102]})
r = df.apply(np.sum) # applies the sum function to all columns
c = r <= 50 #create boolean test for columns
df[c[c].index] #Use boolea indexing to get columns and column filter for dataframe
Output:
A
0 2
1 4
2 6
3 8
Updating an old answer:
df.loc[:, df.sum() <= 50]
I have a large dataframe containing lots of columns.
For each row/index in the dataframe I do some operations, read in some ancilliary ata, etc and get a new value. Is there a way to add that new value into a new column at the correct row/index?
I can use .assign to add a new column but as I'm looping over the rows and only generating the data to add for one value at a time (generating it is quite involved). When it's generated I'd like to immediately add it to the dataframe rather than waiting until I've generated the entire series.
This doesn't work and gives a key error:
df['new_column_name'].iloc[this_row]=value
Do I need to initialise the column first or something?
There are two steps to created & populate a new column using only a row number...
(in this approach iloc is not used)
First, get the row index value by using the row number
rowIndex = df.index[someRowNumber]
Then, use row index with the loc function to reference the specific row and add the new column / value
df.loc[rowIndex, 'New Column Title'] = "some value"
These two steps can be combine into one line as follows
df.loc[df.index[someRowNumber], 'New Column Title'] = "some value"
If you have a dataframe like
import pandas as pd
df = pd.DataFrame(data={'X': [1.5, 6.777, 2.444, pd.np.NaN], 'Y': [1.111, pd.np.NaN, 8.77, pd.np.NaN], 'Z': [5.0, 2.333, 10, 6.6666]})
Instead of iloc,you can use .loc with row index and column name like df.loc[row_indexer,column_indexer]=value
df.loc[[0,3],'Z'] = 3
Output:
X Y Z
0 1.500 1.111 3.000
1 6.777 NaN 2.333
2 2.444 8.770 10.000
3 NaN NaN 3.000
If you want to add values to certain rows in a new column, depending on values in other cells of the dataframe you can do it like this:
import pandas as pd
df = pd.DataFrame(data={"A":[1,1,2,2], "B":[1,2,3,4]})
Add value in a new column based on the values in cloumn "A":
df.loc[df.A == 2, "C"] = 100
This creates the column "C" and addes the value 100 to it, if column "A" is 2.
Output:
A B C
0 1 1 NaN
1 1 2 NaN
2 2 3 100
3 2 4 100
It is not necessary to initialise the column first.
You can just use pandas built in function DataFrame.at
You can chose a list on several index or a single index and column
df.at[4, 'B'] = 10