EDITED
I have a pandas dataframe like so:
data = {'ID': ['A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'A', 'B', 'B', 'B', 'B', 'B', 'B', 'B', 'B', 'B', 'B'],'column_1': [0, 0, 0, 0, 0.1, 1, 1.5, 2, 3, 4, 4.5, 5, 4.9, 3, 2, 1.8, 1, 0, 0, 1, 3, 0, 1.3, 2, 3, 4.3, 4.8, 5, 4.2, 3.5, 3, 2.6, 2, 1.9, 1, 0, 0, 0, 0, 0, 0.1, 0.2, 0.3, 1, 2, 3, 5, 4, 2, 0.5, 0, 0],'column_2': [13, 25, 96, 59, 5, 92, 82, 141, 50, 85, 84, 113, 119, 128, 8, 133, 82, 10, 15, 62, 11, 68, 18, 24, 37, 55, 83, 48, 13, 81, 43, 36, 56, 43, 36, 46, 45, 127, 55, 67, 113, 98, 78, 78, 57, 131, 121, 126, 142, 51, 64, 95]} ID column_1 column_20 A 0.0 131 A 0.0 252 A 0.0 963 A 0.0 594 A 0.1 55 A 1.0 926 A 1.5 827 A 2.0 1418 A 3.0 509 A 4.0 8510 A 4.5 8411 A 5.0 11312 A 4.9 11913 A 3.0 12814 A 2.0 815 A 1.8 13316 A 1.0 8217 A 0.0 1018 A 0.0 1519 A 1.0 6220 A 3.0 1121 A 0.0 6822 A 1.3 1823 A 2.0 2424 A 3.0 3725 A 4.3 5526 A 4.8 8327 A 5.0 4828 A 4.2 1329 A 3.5 8130 A 3.0 4331 A 2.6 3632 A 2.0 5633 A 1.9 4334 A 1.0 3635 A 0.0 4636 A 0.0 4537 A 0.0 12738 A 0.0 5539 A 0.0 6740 A 0.1 11341 A 0.2 9842 B 0.3 7843 B 1.0 7844 B 2.0 5745 B 3.0 13146 B 5.0 12147 B 4.0 12648 B 2.0 14249 B 0.5 5150 B 0.0 6451 B 0.0 95Tracing back from when the value hits 5 in column_1, I want to find the value in column_2 just before the value in column_1 increased from 0 and just after it came back down to 0. So, in the data frame above, the values in column_2 would be 5, 10 and 18, 46.I want to perform some arithmetic and would like to add 2 columns before& after with those values grouped by the ID column.The expected output would be:
ID column_1 column_2 Before After0 A 0.0 13 0 01 A 0.0 25 0 02 A 0.0 96 0 03 A 0.0 59 0 04 A 0.1 5 0 05 A 1.0 92 0 06 A 1.5 82 0 07 A 2.0 141 0 08 A 3.0 50 0 09 A 4.0 85 0 010 A 4.5 84 0 011 A 5.0 113 5 1012 A 4.9 119 0 013 A 3.0 128 0 014 A 2.0 8 0 015 A 1.8 133 0 016 A 1.0 82 0 017 A 0.0 10 0 018 A 0.0 15 0 019 A 1.0 62 0 020 A 3.0 11 0 021 A 0.0 68 0 022 A 1.3 18 0 023 A 2.0 24 0 024 A 3.0 37 0 025 A 4.3 55 0 026 A 4.8 83 0 027 A 5.0 48 18 4628 A 4.2 13 0 029 A 3.5 81 0 030 A 3.0 43 0 031 A 2.6 36 0 032 A 2.0 56 0 033 A 1.9 43 0 034 A 1.0 36 0 035 A 0.0 46 0 036 A 0.0 45 0 037 A 0.0 127 0 038 A 0.0 55 0 039 A 0.0 67 0 040 A 0.1 113 0 041 A 0.2 98 0 042 B 0.3 78 0 043 B 1.0 78 0 044 B 2.0 57 0 045 B 3.0 131 0 046 B 5.0 121 78 6447 B 4.0 126 0 048 B 2.0 142 0 049 B 0.5 51 0 050 B 0.0 64 0 051 B 0.0 95 0 0For a given ID if column_1 starts with a non zero value, it should give the first value of column_2 for that group.
The rest of the rows in Before and After can be filled with null or zeroes.Is there an elegant way to achieve this?
