Inserting String at first index number pandas dataframe

Question

I have a dataframe that looks like this:

  Index Z1  Z2  Z3 Z4
    0   0   0   A  A
    1   0   B   0  0
    2   C   0   C  0

I want to convert it to

 Index  Z1  Z2  Z3 Z4
   ABC 
    0   0   0   A  A
    1   0   B   0  0
    2   C   0   C  0

I want to basically insert text "ABC" before 0th index and the entire row(ABC) will not have any value It will be okay if the index starts from 1 and not 0 as below

 Index  Z1  Z2  Z3 Z4
   ABC 
    1   0   0   A  A
    2   0   B   0  0
    3   C   0   C  0

What is the purpose of this? While we're all keen to contribute with answers, I think this needs careful consideration. Why would you add a string to an integer index? — jpp, Feb 16 '18 at 11:28
Main purpose is that I have many such dataframe which I am appending to create a final csv.So inorder to differentiate between these dataframes in that final csv I am inserting text at the beginning of each dataframe — Rookie_123, Feb 16 '18 at 11:31
@Rookie_123, don't worry, the balance is still positive: `5+5-2-2 = +6` ;) — MaxU - stand with Ukraine, Feb 16 '18 at 11:45
No I was not actually worried about that.My main concern was there should have been some concrete reason for downvoting it. — Rookie_123, Feb 16 '18 at 11:47

jezrael · Accepted Answer · 2018-02-16T11:38:33.330

It seems you need set index name or columns name. But there is possible some pandas function should remove it.

print (df.index.name)
None
print (df.columns.name)
Index

df.index += 1
df.index.name = 'ABC'
print (df)
Index Z1 Z2 Z3 Z4
ABC              
1      0  0  A  A
2      0  B  0  0
3      C  0  C  0

Better is create MultiIndex and select by loc:

mux = pd.MultiIndex.from_product([['ABC'], df.index + 1])
df = df.set_index(mux)
print (df)
      Z1 Z2 Z3 Z4
ABC 1  0  0  A  A
    2  0  B  0  0
    3  C  0  C  0

print (df.loc['ABC'])
  Z1 Z2 Z3 Z4
1  0  0  A  A
2  0  B  0  0
3  C  0  C  0

Also if need distingush many DataFrames with same structure use concat with parameter keys for MultiIndex:

#sample only, in real df1, df2, ... dfn
dfs = [df,df,df]
df = pd.concat(dfs, keys=('a','b','c'))
print (df)
    Z1 Z2 Z3 Z4
a 0  0  0  A  A
  1  0  B  0  0
  2  C  0  C  0
b 0  0  0  A  A
  1  0  B  0  0
  2  C  0  C  0
c 0  0  0  A  A
  1  0  B  0  0
  2  C  0  C  0

MaxU - stand with Ukraine · Answer 2 · 2018-02-16T11:41:04.793

In [53]: df.rename_axis('ABC').rename_axis('Index',1)
Out[53]:
Index Z1 Z2 Z3 Z4
ABC
0      0  0  A  A
1      0  B  0  0
2      C  0  C  0

index starting with 1:

In [54]: df.set_index(df.index+1).rename_axis('ABC').rename_axis('Index',1)
Out[54]:
Index Z1 Z2 Z3 Z4
ABC
1      0  0  A  A
2      0  B  0  0
3      C  0  C  0

UPDATE:

Main purpose is that I have many such dataframe which I am appending to create a final csv.So inorder to differentiate between these dataframes in that final csv I am inserting text at the beginning of each dataframe

I would suggest to store your data in HDF5 file(s) - it has quite a few advantages compared to CSV:

In terms of IO operations (read/write) HDF5 is orders of magnitude faster compared to CSV
it preserves original dtypes (numerical, strings, datetimes, etc.)
it's indexes (you can store data sets that don't fit in memory and conditionally read only a needed subset)
you can store multiple data sets (DataFrame's) in one file using different keys
it supports compression (blosc, lzo, zlib, bzip2)

Demo:

df.to_hdf(r'/path/to/file.h5', 'ABC', data_columns=True)
df2.to_hdf(r'/path/to/file.h5', 'DEF', data_columns=True)

Yes I thought if that was not possible(to start index from 0) even starting index from 1 would be fine.Just for the ease of it — Rookie_123, Feb 16 '18 at 11:32

Inserting String at first index number pandas dataframe

2 Answers2