CODEWITHSUNDEEP
pythonpandasseries
KinWolf
KinWolf

Reputation: 175

How to split series in two columns pandas

I have three lists filled with data and I would like to concatenate them to create a dataframe

type of data_activationsLV : list
type of data_activationsF : list
type of data_activationsPC : list

The structure of the data for the three lists :

data_activationsLV     data_activationsF       data_activationsPC
    index    a         index     b              index     c
    14468    7.8       14468     7.2            14468     7.6         
    14469    7.8       14469     7.1            14469     7.0
    14470    7.9       14470     7.9            14470     8.1
    14471    8.2       14471     9.5            14471     9.9

..

I transform them into series and concate them :

df15LV = pd.Series(data_activationsLV)
df15F = pd.Series(data_activationsF)
df15PC = pd.Series(data_activationsPC)

dfnew2=pd.concat([df15LV,df15F,df15PC], ignore_index=True, axis=1)

by cons here I have a problem, in each column, it considers the name of the old column and its index with the value

index    0              1                2
0        a14468  7.8    b14468   7.2     c14468   7.6
1        a14469  7.8    b14469   7.1     c14469   7.0
2        a14470  7.9    b14470   7.9     c14470   8.1
3        a14471  8.2    b14471   9.5     c14471   9.9

So I tested the split function :

dfnew2['a'] = dfnew2[2].split(' ')

But it did not work,When I try to split those columns, this is what happens:

AttributeError: 'Series' object has no attribute 'split'

Is it possible to have only the values for each columns:

index    df15LV     df15F      df15PC
0        7.8        7.2        7.6
1        7.8        7.1        7.0
2        7.9        7.9        8.1
3        8.2        9.5        9.9

Upvotes: 2

Views: 5517

Answers (2)

jezrael
jezrael

Reputation: 863236

I think you need for splitting apply with str.split and selecting with str[1]:

print (data_activationsLV)
['14468  7.8', '14469  7.8']

print (data_activationsF)
['14468  7.2', '14469  7.1', '14470  7.9', '14471  9.5']

print (data_activationsPC)
['14468  7.6', '14470  8.1', '14471  9.9']

df15LV = pd.Series(data_activationsLV)
df15F = pd.Series(data_activationsF)
df15PC = pd.Series(data_activationsPC)

dfnew2=pd.concat([df15LV,df15F,df15PC], axis=1)
dfnew2 = dfnew2.apply(lambda x: x.str.split().str[1])
#if necessary convert to float
dfnew2 = dfnew2.astype(float)
print (dfnew2)
     0    1    2
0  7.8  7.2  7.6
1  7.8  7.1  8.1
2  NaN  7.9  9.9
3  NaN  9.5  NaN

Another solution is use list comprehension for splitting:

print (data_activationsLV)
['7.8', '7.8']
print (data_activationsF)
['7.2', '7.1', '7.9', '9.5']

print (data_activationsPC)
['7.6', '8.1', '9.9']

df15LV = pd.Series(data_activationsLV)
df15F = pd.Series(data_activationsF)
df15PC = pd.Series(data_activationsPC)

dfnew2=pd.concat([df15LV,df15F,df15PC], axis=1)
#if necessary convert to float
dfnew2 = dfnew2.astype(float)
print (dfnew2)
     0    1    2
0  7.8  7.2  7.6
1  7.8  7.1  8.1
2  NaN  7.9  9.9
3  NaN  9.5  NaN

Upvotes: 2

Niobe
Niobe

Reputation: 1

if you have lists of equal length, you could just create an empty dataframe and fill it:

data_activationsLV = [7.8,7.8,7.9,8.2]
data_activationsF = [7.2,7.1,7.9,9.5]
# create an empty dataframe
columns = ['LV', 'F']
index = np.arange(len(data_activationsLV)) # array of numbers for the number of rows
df = pd.DataFrame(columns=columns, index = index)
df['LV'] = data_activationsLV
df['F'] = data_activationsF
df

Upvotes: 0

Related Questions