如何从 arrays 数组创建 DataFrame 实例？

Question

I have created an array which returns (6, 20) as an attribute of the shape , like this:

import numpy as np

data = np.random.logistic(10, 1, 120)
data = data.reshape(6, 20) 

instantiate pandas.DataFrame from array data

import pandas as pd

data = pd.DataFrame(data)

now this is a dataframe created using data values that come from the numpy module's distributive function

and return this:

          0          1          2          3          4          5   
0   9.602117   9.507674   9.848685   9.215080  11.061676   9.627753   
1  11.702407   9.804924   7.375905  10.784320   8.485818  10.938005   
2   9.628927   9.713187  10.027626  10.653311  11.301493   8.756792   
3  11.229905  12.013172  10.023200   9.211614   7.139757   9.687851   

          6          7         8         9          10         11         12  
0   9.356069  11.483162  8.993130  8.015089   9.808234   9.435853   9.773375   
1  13.422060  10.027434  9.694008  9.677682  10.806266  12.393364   9.479257   
2  10.821846  10.690378  8.321566  9.595122  11.753948  10.021815  10.412572   
3   8.499120   7.352394  9.288662  9.178306  10.073842   9.246110   9.075350   

          13         14         15         16         17         18         19  
0   9.809366   8.502451  11.624395  12.824338   9.729167   8.945258  10.464157  
1   6.698941   9.416421  11.477242   9.622115   6.374589   9.459355  10.435674  
2  11.068721   9.775433   9.447799   8.972052  10.692942  10.978305  10.047067  
3  10.381596  10.968330  11.892766  12.241880   9.980124   7.321942   9.241030

when I try to set columns=list("abcdef") , I get this error:

ValueError: Shape of passed values is (6, 20), indices imply (6, 6)

and my expected output is similar to that shown directly from the numpy array. It should contain each column as a pandas.Series of lists (or list of lists).

                       a.  
0 [ 6.98467276  9.16242742  6.99065177 11.50834399  9.29697138  7.93926441
   9.05857668  7.13652948 11.01724792 13.31658877  8.63137079  9.5564405
   7.37161153 11.19414704  9.45957466  9.19826796 10.13506672  9.74830158
   9.97456348  8.35217153] 
                        b.
 [10.48249082 11.94030324 12.59080011 10.55695088 12.43071037 11.49568774
  10.03540181 11.08708832 10.24655111  8.17904856 11.04791142  7.30069964
   8.34783674  9.93743588  8.1537666   9.92773204 10.3416315   9.51624921
   9.60124236 11.37511301] 
                        c.
 [ 8.21851024 12.71641524  9.7748047   9.51267978  7.92793378 12.1646706
   9.67236267 10.22201002  9.67197374  9.70551429  7.79209516  9.20295594
   9.26231527  8.04560836 11.0409066   8.63660332  9.18397671  8.17510874
   9.61619671  8.42704322] 
                        d.
 [14.54825819 16.97573893  7.70643136 12.06334323 14.64054726  9.54619595
  10.30686621 12.20487566 10.78492189 12.01011666 10.12405213  8.57057999
  10.41665479  7.85921253 10.15572125  9.20554292 10.03832545  9.43720211
  11.06605713  9.60298514]

I have found this thread that looks like my problem but it has not helped me much, also I would use the data in a different way.

Could I assign the lengths of the columns or maybe assign the dimensions of this Pandas.DataFrame ?

Answer 1

Your data has 6 rows and 20 columns. If you want to pass each "row" of the numpy array as a "column" to the DataFrame, you can simply transpose :

df = pd.DataFrame(data=np.random.logistic(10, 1, 120).reshape(6,20).transpose(), 
                  columns=list("abcdef"))

Edit:

To get the data in a single row, try:

df = pd.DataFrame(columns=list("abcdef"), index=[0])
df.iloc[0] = np.random.logistic(10, 1, 120).reshape(6,20).transpose()