My dataframe inside a column data is [1,1,2,3,4,7,8,8,15,19,20,21]. I want to get the most contiguous data segments in this column: [1,2,3,4]. How to calculate it?

CodePudding user response：

You can create groups by consecutive values by compare difference with cumulative sum, get counts by GroupBy.transform and last filter maximal counts of original column col - output are all consecutive values with maximal counts:

s = df['col'].groupby(df['col'].diff().ne(1).cumsum()).transform('size')

out = df.loc[s.eq(s.max()), 'col']

If need first maximum consecutives values use Series.value_counts with Series.idxmax:

s = df['col'].diff().ne(1).cumsum()

out = df.loc[s.eq(s.value_counts().idxmax()), 'col']