two dataframe fet difference ( numpy )

Question

d1 = df1.filter(items=['col_a']) # len(d1) = 900
d2 = df.filter(items=['col_a']) # len(d2) = 920

d1 ---------->
    col_a
0   802584
1   408852
2   408311
3   409633
4   801848
5   123123
6   232333

d2 ---------->
    col_a
0   802584
1   408852
2   408311
3   409633
4   801848

col_a is integer value and I want to compare between d1 and d2 to extract differences.

How can I get a new dataframe that includes only 20 values for col_a?

In this situation, I want to get

new_df ----->
     col_a
0    123123
1    232333

Hi @Daniel Unfortunately, your question is not clear enough, you can add Input and expected outputs as well to clarify. — Fading Origami, Nov 10 '20 at 05:24
@FadingOrigami Thank you for concerning, I updated what I exactly want to get — Daniel Lee, Nov 10 '20 at 05:28
Does this answer your question? [Compare two DataFrames and output their differences side-by-side](https://stackoverflow.com/questions/17095101/compare-two-dataframes-and-output-their-differences-side-by-side) — Inyoung Kim 김인영, Nov 10 '20 at 05:33
Try answer from [this](https://stackoverflow.com/a/28902170/8471995) question. `df1[~df1.isin(df2)].dropna(how = 'all')` — Inyoung Kim 김인영, Nov 10 '20 at 05:57

score 2 · Accepted Answer · edited Nov 10 '20 at 21:03

2

You can do this using df.isin and negation (~):

In [1278]: d1[~d1.col_a.isin(d2.col_a)]
Out[1278]: 
    col_a
5  123123
6  232333

edited Nov 10 '20 at 21:03

halfer

19,824
17
99
186

answered Nov 10 '20 at 06:01

Mayank Porwal

33,470
8
37
58

A question. Can `negation (~)` all python? Or is it `pandas` only? – Inyoung Kim 김인영 Nov 10 '20 at 06:04
Python `not` operator is `!`. Pandas `~` is mostly like `not in` operator in `SQL`. – Mayank Porwal Nov 10 '20 at 06:06
it shows empty in my code but looks correct in your code. both type integer64. I'll check data. @InyoungKim Your answer shows same like d1 – Daniel Lee Nov 10 '20 at 06:06
I tested my code on the sample you provided. Works fine for me. Please make sure you are running my exact command. – Mayank Porwal Nov 10 '20 at 06:08
Okay. If you proved I'll use your code to test my data. Thank you so much – Daniel Lee Nov 10 '20 at 06:10

two dataframe fet difference ( numpy )

1 Answers1