CODEWITHSUNDEEP
pythoncsvpandasdataframesave
Federico Gentile
Federico Gentile

Reputation: 5940

Saving dataframe and additional data in same .csv file with Pandas

Description: I have a set of parameters (par1, par2, par3, par4) and a dataframe df. In this example the parameters and the number of columns of the dataframe are respectively are 4 and 3 but they could both be a generic number. 

import pandas as pd
import numpy as np

# list of parameters
par1 = 1.05
par2 = 20
par3 = 50000
par4 = 12315468

# Dataframe
dic = {'A' : ['PINCO','PALLO','TOLLO','FINGO','VOLVA'],
       'B' : [ 4     , 5     , np.nan, 1     , 0],
       'C' : [ 1     , 4     , 8     , 7     , 6]}
df = pd.DataFrame(dic)

My goal is to save this data in the same .csv file but I don't know how to do it since the number of parameters doesn't match the number of columns of df.

My output file must should follow this rule:

Therefore it looks like this:

enter image description here

Question: Could you provide a smart and efficient way to obtain the output file with the desired shape?

Upvotes: 2

Views: 3381

Answers (2)

jezrael
jezrael

Reputation: 863166

You can first create list of parameters pars, then overwrite columns by pars with same length and last use reindex, but values has to be unique:

# list of parameters
par1 = 1.05
par2 = 20
par3 = 50000
par4 = 12315468

pars = [par1,par2,par3,par4]

# Dataframe
dic = {'A' : ['PINCO','PALLO','TOLLO','FINGO','VOLVA'],
       'B' : [ 4     , 5     , np.nan, 1     , 0],
       'C' : [ 1     , 4     , 8     , 7     , 6]}
df = pd.DataFrame(dic)

df.columns = pars[:len(pars) - 1]
print (df)
  1.05      20.00     50000.00
0    PINCO       4.0         1
1    PALLO       5.0         4
2    TOLLO       NaN         8
3    FINGO       1.0         7
4    VOLVA       0.0         6

df = df.reindex(columns=pars)
print (df)
  1.05         20.00        50000.00     12315468.00
0       PINCO          4.0            1          NaN
1       PALLO          5.0            4          NaN
2       TOLLO          NaN            8          NaN
3       FINGO          1.0            7          NaN
4       VOLVA          0.0            6          NaN

Another possible solution is use concat of DataFrame created from list pars:

pars = [par1,par2,par3,par4]

# Dataframe
dic = {'A' : ['PINCO','PALLO','TOLLO','FINGO','VOLVA'],
       'B' : [ 4     , 5     , np.nan, 1     , 0],
       'C' : [ 1     , 4     , 8     , 7     , 6]}
df = pd.DataFrame(dic)
print (df)

df.columns = range(len(df.columns))
s = pd.DataFrame([pars])
print (s)
      0   1      2         3
0  1.05  20  50000  12315468

df1 = pd.concat([s, df], ignore_index=True)
print (df1)
       0     1      2           3
0   1.05  20.0  50000  12315468.0
1  PINCO   4.0      1         NaN
2  PALLO   5.0      4         NaN
3  TOLLO   NaN      8         NaN
4  FINGO   1.0      7         NaN
5  VOLVA   0.0      6         NaN

EDIT Also is possible use mode a for append in read_csv:

filename = 'filename.csv'
pars = [par1,par2,par3,par4]
pd.DataFrame([pars]).to_csv(filename, index=False, header=False)
df.to_csv(filename, index=False, header=False, mode='a')

Upvotes: 2

Shreyash S Sarnayak
Shreyash S Sarnayak

Reputation: 2335

First create list of parameters par1, par2, par3, etc,.

l = [par1, par2, par3, par4]

Then save the list to csv

filename = 'abc.csv'
pd.DataFrame(l).T.to_csv(filename, index=False, header=False)

Use python's append mode to append the data frame to csv.

with open(filename, 'a') as f:  # Use append mode.
    df.to_csv(f, index=False, header=False)

Upvotes: 4

Related Questions