如何使用pandas数据框的列值更改numpy数组的索引值

Question

I have a dataframe: data:

  user_id   item_id rating
0    772       36    3
1    471      228    5
2    641      401    4
3    312       98    4
4     58      504    5

I have created a array of random number, whose size equal to unique values in item_id

a = np.random.random(1662)

Now I want to convert index value of array as column value of item_id. How to change index of array. Item_id has total 1662 value but it is not continues like 1 to 1662. there are some values missing. the maximum value being 1681. So what I need is to create a random array of size 1662(ie equal to unique values of item_id). But if i want to see random value associated with a index (1677) how can i see that?

Since size of array is 1662 so the maximum index value is 1661 for index. so i want to associate 1662 number with the values of item_id

Answer 1

Your question seems vague.

You have a random array with 1662 unique values in item_id - meaning that its possible a value exists in item_id and its also possible that it doesn't. If it does, you want to re-index the numpy array a according to the value - so if the number 36 is in 'a' and in 'item_id' you want a[36] = 36?

Please clarify.

If that's what you want, then simply create an ordered list of every item_id like so:

a = np.arrange(max(data.item_id))
b = np.isin(a, data.item_id)
a[b] = np.full(a[b], data.item_id)

UPDATE: For a[36] = random-number

a[b] = np.full(a[b], np.random.normal())

Suggestion : You can also add random numbers inside your dataframe as:

data['random'] = np.random.random(data.count())
# create a dict for indexing
random_dict = dict(zip(data['item_id'].tolist(), data['random'].tolist()))
# or set item_id as index:
data.index = data['item_id'].values

Let me know if this helps!