In the data frame of probabilities over time return first column name where value is < .5 for each row

Given a pandas data frame like the following where the column names are the time, the rows are each of the subjects, and the values are probabilities return the column name (or time) the first time the probability is less than .50 for each subject in the data frame. The probabilities are always descending from 1-0 I. have tried looping though the data frame but it is not computationally efficient.

subject id	0	1	2	3	4	5	6	7	…	669	670	671
1	1	0.997913	0.993116	0.989017	0.976157	0.973078	0.968056	0.963685	…	0.156092	0.156092	0.156092
2	1	0.990335	0.988685	0.983145	0.964912	0.958	0.952	0.946995	…	0.148434	0.148434	0.148434
3	1	0.996231	0.990571	0.985775	0.976809	0.972736	0.969633	0.966116	…	0.17037	0.17037	0.17037
4	1	0.997129	0.994417	0.991054	0.978795	0.974216	0.96806	0.963039	…	0.15192	0.15192	0.15192
5	1	0.997728	0.993598	0.986641	0.98246	0.977371	0.972874	0.96816	…	0.154545	0.154545	0.154545
6	1	0.998134	0.995564	0.989901	0.986941	0.982313	0.972951	0.969645	…	0.17473	0.17473	0.17473
7	1	0.995681	0.994131	0.990401	0.974494	0.967941	0.961859	0.956636	…	0.144753	0.144753	0.144753
8	1	0.997541	0.994904	0.991941	0.983389	0.979375	0.973158	0.966358	…	0.158763	0.158763	0.158763
9	1	0.992253	0.989064	0.979258	0.955747	0.948842	0.942899	0.935784	…	0.150291	0.150291	0.150291

Goal Output

subject id	time prob < .05
1	100
2	99
3	34
4	19
5	600
6	500
7	222
8	111
9	332

https://stackoverflow.com/questions/66837503/in-the-data-frame-of-probabilities-over-time-return-first-column-name-where-valu March 28, 2021 at 08:52AM