i have pandas dataframe:
import numpy import pandas df1 = abs((pandas.dataframe(numpy.random.randn(20, 8))*100).astype(int)) df1.columns = list('abcdefgh') df1.index = pandas.date_range('1/1/2014', periods=20) 
how create new column give me minimum value of first half of current row , last 3 values in previous row?
for example, first 5 rows in created column be:
nan 12 4 14 21
here 1 way it. basically, need first shift last 3 columns , combine first 4 columns, , calculate min.
import numpy import pandas # data # =================================== numpy.random.seed(0) df1 = abs((pandas.dataframe(numpy.random.randn(20, 8))*100).astype(int)) df1.columns = list('abcdefgh') df1.index = pandas.date_range('1/1/2014', periods=20) # processing # =================================== df1['custom_min'] = pandas.concat([df1[df1.columns[:5]], df1[df1.columns[-3:]].shift(1)], axis=1).min(axis=1) print(df1) b c d e f g h custom_min 2014-01-01 176 40 97 224 186 97 95 15 40 2014-01-02 10 41 14 145 76 12 44 33 10 2014-01-03 149 20 31 85 255 65 86 74 12 2014-01-04 226 145 4 18 153 146 15 37 4 2014-01-05 88 198 34 15 123 120 38 30 15 2014-01-06 104 142 170 195 50 43 125 77 30 2014-01-07 161 21 89 38 51 118 2 42 21 2014-01-08 6 30 63 36 67 35 81 172 2 2014-01-09 17 40 163 46 90 5 72 12 17 2014-01-10 113 123 40 68 87 57 31 5 5 2014-01-11 116 90 46 153 148 189 117 17 5 2014-01-12 107 105 40 122 20 97 35 70 17 2014-01-13 1 178 12 40 188 134 127 96 1 2014-01-14 117 194 41 74 192 148 186 90 41 2014-01-15 86 191 26 80 94 15 61 92 26 2014-01-16 37 109 29 132 69 14 43 184 15 2014-01-17 67 40 76 53 67 3 63 67 14 2014-01-18 57 20 39 109 149 43 16 63 3 2014-01-19 238 94 91 111 131 46 6 171 16 2014-01-20 74 82 9 66 112 107 114 43 6 
Comments
Post a Comment