Extract rows from dataframe based on values in list

Question

So I have a dataframe that looks like this:

split_df = 
     filename  width  height  class  xmin  ymin  xmax  ymax
0  128778.JPG   5472    3648  Panel  4489   818  4915  1417
1  128778.JPG   5472    3648  Panel  3023  1278  3403  1854
2  128776.JPG   5472    3648  Panel  3369  1310  3759  1890
3  128776.JPG   5472    3648  Panel  3721  1339  4116  1919
4  128775.JPG   5472    3648  Panel  4076  1368  4476  1952

And then two lists, train_set and test_set, these are two lists like so:

train_set = ["128778.JPG", "128776.JPG"]
test_set = ["128775.JPG"]

So basically I was wondering if there's a way to split the dataframe based on these two lists kinda like this:

> train_df = split(split_df, train_set)
> train_df 
>        filename  width  height  class  xmin  ymin  xmax  ymax
    0  128778.JPG   5472    3648  Panel  4489   818  4915  1417
    1  128778.JPG   5472    3648  Panel  3023  1278  3403  1854
    2  128776.JPG   5472    3648  Panel  3369  1310  3759  1890
    3  128776.JPG   5472    3648  Panel  3721  1339  4116  1919
> test_df = split(split_df, test_set)
> test_df 

>         filename  width  height  class  xmin  ymin  xmax  ymax
    0   128775.JPG   5472    3648  Panel  4076  1368  4476  1952

Yep, since your hypothetical `split` function is called each time, saving no computation, this is nothing but a basic filter. — roganjosh, Jan 15 '19 at 22:53
I haven't really worked a lot with pandas, I just used split as a stand-in for some built in method. — PEREZje, Jan 15 '19 at 22:54
@W-B `!= "128775.JPG"` would do in this case. I suspect it's faster than `isin`? The OP intends to do the separation in 2 parts already. There's a chronic lack of effort here. — roganjosh, Jan 15 '19 at 22:55
@roganjosh this is base on the assumption there are no other corrupt lines in the original dataframe other than above three filenames — BENY, Jan 15 '19 at 22:56

Extract rows from dataframe based on values in list

0 Answers0