CODEWITHSUNDEEP
pythonpandas
Viktor.w
Viktor.w

Reputation: 2297

Pandas: after using qcut(data,3), how to find the range of the quantile

My data looks like this:

    spread                              CPB%    Bin  
0  0.00000787  0.001270648030495552731893265565   B  
1  0.00000785  0.003821656050955414012738853503   A  
2  0.00000749  0.005821656050955414012738853503   C  
3  0.00000788  0.004821656050955414012738853503   B

So I have basically assigned a letter A,B or C according to the value of their spread. I have done this using this code:

s = (df['spread'] * 10**15).astype(np.int64)
df['Bin'] = pd.qcut(s, 3, labels=list('ABC'))

What I need to do now, is that I have 100 spreads (from 0.000001 to 0.0001) and I need to know if they end-up in the Bin A,B or C. Is there a way to find let's say the 'range' of the above quantile?

More precisely I have the below spreads:

      spread
0   0.000100
1   0.000109
2   0.000118
3   0.000127
4   0.000136
5   0.000145

How can I know if they end-up in the same bin as A-B-C of above? Thanks

Upvotes: 1

Views: 2723

Answers (2)

StackerDekker
StackerDekker

Reputation: 71

If you use:

df['bins'] = pd.qcut(df['your_split_col_name'], 3)

The output will tell you the bin intervals.

Using the labels masks that.

EDIT

To use the created split, you could use pandas groupby function.

df['bins'] = pd.qcut(df['your_split_col_name'], 3)
df = df.groupby('bins')
df.describe()

Upvotes: 1

jezrael
jezrael

Reputation: 862791

I believe you need add parameter retbins=True for qcut for return intervals, so is possible reuse it in cut:

print (df1)
     spread      CPB% Bin
0  0.000008  0.001271   B
1  0.000008  0.003822   A
2  0.000007  0.005822   C
3  0.000008  0.004822   B

print (df2)
     spread
0  0.000008 <-change data sample for match
1  0.000109
2  0.000118
3  0.000127
4  0.000136
5  0.000145

s = (df1['spread'] * 10**15).astype(np.int64)
v,b = pd.qcut(s, 3, labels=list('ABC'),retbins=True)
print (v)
0    B
1    A
2    A
3    C
Name: spread, dtype: category
Categories (3, object): [A < B < C]

print (b)
[7490000000 7849999999 7869999999 7880000000]


s1 = (df2['spread'] * 10**15).astype(np.int64)
df2['new'] = pd.cut(s1, bins=b, labels=v.cat.categories)
print (df2)
     spread  new
0  0.000008    A
1  0.000109  NaN
2  0.000118  NaN
3  0.000127  NaN
4  0.000136  NaN
5  0.000145  NaN

Upvotes: 2

Related Questions