如何计算 Pandas 中不同唯一 ID 的特定值出现的次数？

Question

I have large dataset of over 10,000 entries. The dataset contains a unique ID, a year that an event occurred, and the size of that event. I want to count the number of events of above and below a specific threshold value for each unique ID. However, for events below the threshold, I only want to count the event if it occurred after a certain year.

As an example, let's say I have the below data:

Unique ID, Year, Size  
111, 1980, 1  
111, 1992, 2  
111, 2000, 4  
222, 1990, 5  
222, 1994, 3  
333, 1999, 2  
333, 2011, 5  
333, 2012, 2  
333, 2016, 1 

I want to categorize how many events are either equal to or above size 3 for each unique ID. But I also only want to count events that are <=3 if they occurred after a specific year. For example, I only want to count events that occurred after 1980 for Unique ID 1, after 1992 for Unique ID 2, and after 2000 for unique ID 3.

Based on the above example data, I would be expecting the following result

Unique ID, <=3, >3

111, 1, 1    
222, 1, 1  
333, 2, 1 

Answer 1

Because there is for each Unique ID different threshold year create dictionary for Series.map , so is possible filter, here by Series.lt for less method with boolean indexing :

d = {111:1980, 222:1992, 333:2000}
df = df[df['Unique ID'].map(d).lt(df['Year'])]
print (df)
   Unique ID  Year  Size
1        111  1992     2
2        111  2000     4
4        222  1994     3
6        333  2011     5
7        333  2012     2
8        333  2016     1

And then for count is used crosstab with numpy.where :

df = pd.crosstab(df['Unique ID'], np.where(df['Size'].le(3), '<=3','>3'))
print (df)
col_0      <=3  >3
Unique ID         
111          1   1
222          1   0
333          2   1