pandas.read_csv 的“千”和“skip_blank_lines”arguments 将无法正常工作。 为什么？

Question

This my code:

in[0]
      import pandas as pd
      df = pd.read_csv('datefile6.csv',thousands=',', skip_blank_lines=True)
      df

out[1]      month   day     year    salary   age
       0    8.0     15.0    2012.0  1400.0   25.0
       1    NaN     NaN     NaN     NaN      NaN
       2    9.0     4.0     2020.0  2500.0   26.0

As we see the thousands did not work. Also, line[1] which is blank has not been skipped by the commands.

I expected ',' from the "thousands" command and removed line[1] from the "skip_blan_lines" command.

Answer 1

thousands parameter is a property of the input file. It tells pandas that numbers in your csv file contain thousands character (typically comma or dot). Parameter thousands does not impact the output.

Consider this code:

import pandas as pd
df = pd.read_csv('datafile6.csv', sep=';', thousands=',', skip_blank_lines=True)
print(df)

where datafile6.csv is:

month;day;year;salary;age
8;15;2,012;1,400;25
9;4;2,020;2,500;26

I get the output as:

   month   day    year  salary   age
0    8.0  15.0  2012.0  1400.0  25.0
1    9.0   4.0  2020.0  2500.0  26.0

and you can see that 1,400 has been correctly parsed as 1400 etc.

Regarding your question about skip_blank_lines. I suspect that instead of a completely blank line, your csv contains field separators.
Consider now this as the content from datafile6.csv:

month;day;year;salary;age
8;15;2,012;1,400;25
;;;;
9;4;2,020;2,500;26

9;3;2,021;3,200;33

I get dataframe output as:

   month   day    year  salary   age
0    8.0  15.0  2012.0  1400.0  25.0
1    NaN   NaN     NaN     NaN   NaN
2    9.0   4.0  2020.0  2500.0  26.0
3    9.0   3.0  2021.0  3200.0  33.0

NaN results from line #3 of datafile6.csv which is not really blank but has 4 field separators. Where as line #5 which is completely blank is skipped. This is the behaviour of parameter skip_blank_lines

Hope this clears.