用固定字符串替换Pandas数据透视表非空结果单元格

Question

我正在尝试将看起来像第一个示例的CSV转换为看起来像下面第二个示例的CSV。

我一直在和Pandas一起玩，并认为我的基础工作正常，但我似乎无法弄明白如何进行最后一次转换（从我在枢轴中的占位符值到实际的英文单词）

在下面的代码中，我需要帮助的部分是注释“我需要弄清楚我可以放在这里的东西，它将替换在列的单元格中找到的任何非空值[c ]使用字符串'registered'。“

注意 - 如果您建议一种更有效的方式来遍历数据而不是列名列表上的for循环，请随意。 for循环只是一种测试功能的方法，因为我第一次使用Pandas。

输入：

First  Last  Email      Program
john   doe   jd@me.com  BasketWeaving
jane   doe   dj@me.com  BasketWeaving
jane   doe   dj@me.com  Acrobatics
jane   doe   dj@me.com  BasketWeaving
mick   jag   mj@me.com  StageDiving

期望的输出：

First  Last  Email      StatusBasketWeaving__c  StatusAcrobatics__c  StatusStageDiving__c
john   doe   jd@me.com  registered
jane   doe   dj@me.com  registered              registered
mick   jag   mj@me.com                                               registered

（实际上我的代码插入了一列，但是这个例子太宽了，所以这里没有显示。）

这是我到目前为止所写的内容：

import pandas
import numpy

# Read in the First Name, Last Name, Email Address, & "Program Registered For" columns of a log file of registrations conducted that day.
tally = pandas.read_csv('tally.csv', names=['First', 'Last', 'Email', 'Program'])

# Rename the First Name & Last Name columns so that they're Salesforce Contact object field names
tally.rename(columns={'First':'FirstName', 'Last':'LastName'}, inplace=True)

# Create a concatenation of First, Last, & Email that can be used for later Excel-based VLOOKUP-ing Salesforce Contact Ids from a daily export of Id+Calculated_Lastname_Firstname_Email from Salesforce
tally['Calculated_Lastname_Firstname_Email__c'] = tally['LastName'] + tally['FirstName'] + tally['Email']

# Rename the values in Program so that they're ready to become field names for the Salesforce Contact object
tally['Program'] = 'Status' + tally['Program'] + '__c'

# Pivot the data by grouping on First+Last+Email+(Concatenated), listing the old registered-for-Program values as column headings, and putting
# a non-null value under that column heading if the person has any rows indicating that they registered for it.
pivottally = pandas.pivot_table(tally, rows=['FirstName', 'LastName', 'Email', 'Calculated_Lastname_Firstname_Email__c'], cols='Program', aggfunc=numpy.size)

# Grab a list of column names that have to do with the programs themselves (these are where we'll want to replace our non-null placeholder with 'Registered')
statuscolumns = [s for s in (list(pivottally.columns.values)) if s.startswith('Status')]

for c in statuscolumns:
    #pivottally.rename(columns={c:'Hi'+c}, inplace=True) # Just a test line to make sure my for loop worked.
    # I need to figure out something I can put here that will replace any non-null value found in the cells of column pivottally[c] with the string 'Registered'

print(pivottally.head())

#pivottally.to_csv('pivottally.csv')

感谢您的帮助。

Answer 1

简单的选择可以完成这项工作。构建列列表并迭代它是没用的，因为所有列都是关注的。其他列在索引中。

pivottally[pandas.notnull(pivottally)] = 'registered'

以下是结果的屏幕截图。