如何找到连续下降的次数(增加) [英] how to find number of consecutive decreases(increases)
本文介绍了如何找到连续下降的次数(增加)的处理方法，对大家解决问题具有一定的参考价值，需要的朋友们下面随着小编来一起学习吧！
问题描述
如何找到连续减少(增加)的数量
我有一个数据框，其中包含500K行和12列(用于月份)，并包含开始和结束月份.每列代表一个月.我需要比较范围(startMonth，endmonth)中的第i个月和第(i + 1)个月的每一行. (注:范围不是恒定的.Every行具有不同的范围大小.)
条件:如果开始月>结束月，我应该看到"Neg99 = -999"
这是我的示例数据:
import pandas as pd
import numpy as np
idx = [1001,1002,1003,1004,1005,1006,1007,1008,1009,1010,1011,1012,1013,1014,1015,1016,1017,1018]
data = {'M_1': [3, 1, 0, 0, 1, 0, 1, 1, 1, 0, 6, 6, 6,0,0,2,0,2],
'M_2': [2, 2, 3, 1, 1, 0, 1, 2, 0, 1, 5, 5, 5,1,1,1,1,2],
'M_3': [1, 3, 2, 2, 1, 0, 1, 2, 1, 0, 4, 4, 4,1,1,0,2,2],
'M_4': [0, 4, 1, 3, 1, 0, 1, 2, 0, 1, 3, 0, 3,1,1,0,0,0],
'M_5': [1, 0, 0, 4, 2, 0, 1, 3, 1, 0, 2, 1, 2,1,1,0,0,0],
'M_6': [2, 0, 0, 0, 3, 0, 1, 3, 0, 1, 1, 2, 1,1,1,0,0,0],
'M_7': [3, 0, 0, 0, 2, 0, 1, 2, 1, 0, 0, 3, 0,0,1,0,0,0],
'M_8': [0, 1, 0, 0, 2, 0, 1, 2, 0, 1, 1, 1, 1,0,0,0,0,0],
'M_9': [0, 2, 0, 0, 1, 0, 1, 2, 1, 0, 0, 0, 0,0,0,0,0,0],
'M_10': [0, 3, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0,0,0,0,0,0],
'M_11': [0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0,0,0,0,0,0],
'M_12': [0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0,0,0,0,0,0]}
startMonth = pd.DataFrame([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 4, 4, 5,1,1,1,1,1],
columns=['start'],index=idx)
endMonth = pd.DataFrame([12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 2,12,12,2,2,2],
columns=['end'], index=idx)
df1 = pd.DataFrame(data, index=idx)
Neg99 = -999
我为日期范围写了bool数组；
arr_bool = (np.less_equal.outer(startMonth.start, range(1,13))
& np.greater_equal.outer(endMonth.end, range(1,13))
)
masked=df1.filter(regex='M_[0-9]').mask(~arr_bool)
我需要找到每行的连续减少和增加.
- 这是减少代码；
# Consecutive Decreases
decr = (np.diff(np.hstack((masked.values, np.zeros((masked.values.shape[0], 1)))), axis=1) > 0).argmin(axis=1)
final_decr = pd.DataFrame(decr,
index=idx, columns=['decr'])
final_decr.decr= np.select( condlist = [startMonth.start > endMonth.end],
choicelist = [Neg99],
default = final_decr.decr)
- 这里是增加代码；
incr = (np.diff(np.hstack((masked.values, np.zeros((masked.values.shape[0], 1)))), axis=1) < 0).argmin(axis=1)
final_incr = pd.DataFrame(incr,
index=idx, columns=['incr'])
final_incr.incr= np.select( condlist = [startMonth.start > endMonth.end],
choicelist = [Neg99],
default = final_incr.incr)
最后，我的预期输出是；
Final increase table (.csv);
idx,my_results,expected_result
1001,0,0
1002,3,3
1003,1,1
1004,4,4
1005,0,0
1006,0,0
1007,0,0
1008,1,1
1009,0,0
1010,1,1
1011,0,3
1012,0,0
1013,-999,-999
1014,5,1
1015,6,1
1016,0,0
1017,0,2
1018,0,0
Final decrease table (.csv);
idx,my_result,expected_result
1001,3,3
1002,0,0
1003,0,0
1004,0,0
1005,0,0
1006,0,0
1007,0,0
1008,0,0
1009,1,1
1010,0,0
1011,0,0
1012,0,3
1013,-999,-999
1014,0,0
1015,0,0
1016,0,2
1017,0,0
1018,0,0
Final NoChange table (.csv);
idx,my_result,expected_result
1001,0,0
1002,0,0
1003,0,0
1004,0,0
1005,3,3
1006,11,11
1007,11,11
1008,0,0
1009,0,0
1010,0,0
1011,0,0
1012,0,0
1013,-999,-999
1014,0,0
1015,0,0
1016,0,0
1017,0,0
1018,2,0
Thanks for your advice!
解决方案
# Consecutive Decreases
decr = (np.diff(np.hstack((masked.values, np.zeros((masked.values.shape[0], 1)))), axis=1) > 0).argmin(axis=1)
final_decr = pd.DataFrame(decr,
index=idx, columns=['decr'])
final_decr.decr= np.select( condlist = [startMonth.start > endMonth.end],
choicelist = [Neg99],
default = final_decr.decr)
- Here is increasees code;
incr = (np.diff(np.hstack((masked.values, np.zeros((masked.values.shape[0], 1)))), axis=1) < 0).argmin(axis=1)
final_incr = pd.DataFrame(incr,
index=idx, columns=['incr'])
final_incr.incr= np.select( condlist = [startMonth.start > endMonth.end],
choicelist = [Neg99],
default = final_incr.incr)
And finally, My expected outputs are;
Final increase table (.csv);
idx,my_results,expected_result
1001,0,0
1002,3,3
1003,1,1
1004,4,4
1005,0,0
1006,0,0
1007,0,0
1008,1,1
1009,0,0
1010,1,1
1011,0,3
1012,0,0
1013,-999,-999
1014,5,1
1015,6,1
1016,0,0
1017,0,2
1018,0,0
Final decrease table (.csv);
idx,my_result,expected_result
1001,3,3
1002,0,0
1003,0,0
1004,0,0
1005,0,0
1006,0,0
1007,0,0
1008,0,0
1009,1,1
1010,0,0
1011,0,0
1012,0,3
1013,-999,-999
1014,0,0
1015,0,0
1016,0,2
1017,0,0
1018,0,0
Final NoChange table (.csv);
idx,my_result,expected_result
1001,0,0
1002,0,0
1003,0,0
1004,0,0
1005,3,3
1006,11,11
1007,11,11
1008,0,0
1009,0,0
1010,0,0
1011,0,0
1012,0,0
1013,-999,-999
1014,0,0
1015,0,0
1016,0,0
1017,0,0
1018,2,0
Thanks for your advice!
解决方案
So I think instead of using argmin, you can use idxmin after renaming the column to get an integer value of the position. Then remove the value of the startMonth such as:
incr = (df1.rename(columns={col:int(col.split('_')[1]) for col in masked.columns})
.diff(-1, axis=1) < 0).mask(~arr_bool).idxmin(axis=1) - startMonth.start
decr = (df1.rename(columns={col:int(col.split('_')[1]) for col in masked.columns})
.diff(-1, axis=1) > 0).mask(~arr_bool).idxmin(axis=1) - startMonth.start
Then you can do the np.select as you do, or probably just a .illna(-999) should be enough as I think with this solution everywhere you have nan in the result will be where your condition startMonth.start > endMonth.end is met
这篇关于如何找到连续下降的次数(增加)的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
查看全文
