pandas to dict: to_dict 不存储所有值

Question

I have a dataframe df with 40000 rows:

              0  bin
0      4.506840  4-5
1      4.506840  4-5
2      4.444245  4-5
3      4.485975  4-5
4      4.527705  4-5
...         ...  ...
39995  6.572475  6-7
39996  6.697665  6-7
39997  6.322095  6-7
39998  6.322095  6-7
39999  6.676800  6-7

It stores for every number in column '0' the interval (bin) it belongs to. I want to convert it to a dict by:

dict(zip(df[0],df.bin))

to get an output like:

{4.506840: '4-5', 4.506840: '4-5', 4.444245: '4-5, ... }

so I want to store every value from '0' and the bin it belongs to. Somehow my dict has a length of 340, not 40000, so it doesn't store all of the rows. My question is: why? And how do I get all 40000 rows in the dict? Cheers!

Answer 1

Perhaps you have multiple values in that column that are the same, and altough python allows you to store multiple values that are the same, you can't store more than one of the same key. I would suggest either you don't use a dict or try to put some identifier on each duplicate in the df before converting it into a dict to distinguish it from other keys. Another way would be to divide the df where the duplicates are or store all the duplicates in a list, although I'm not sure this is what you desire.

Answer 2

Due to the duplicates you have in your df[0] , and due to the fact that you cannot have the same key duplicated in a python dictionary, you can do:

result = {}
for i_0, i_bin in zip(df[0],df.bin):
    if i_0 not in result.keys():
        result[i_0] = []
    result[i_0].append(i_bin)

output:

{
    "4.506840": ["4-5", "4-5"],
    "4.444245": ["4-5"],
    ...
}

It depends on what you want to achieve, but this is a way to perceive all the values.

Edit:

As per @anky comment, you can make use of pandas aggregation function to do the same instead of the loop. Definitely, it is of better performance:

df.groupby(0)['bin'].agg(list).to_dict()