Question

我试图重新格式化CSV以将每个月的列转换为每个记录的一个单独的行（基本上是旋转它），即：

分为：

要做到这一点，我认为最好的方法是：

遍历每一行，遍历每个月的列（Jan-17，Feb-17等等），然后复制该行。
然后将月份和总计插入Date和Totals列。
然后删除那个重复的记录并从索引停止的地方开始（即，在每个日期的5个记录循环之后，开始的索引将是5）。
然后，当所有行都重复时，请删除月份列（Jan-17，Feb-17等...）

它用于第一个数据行（即brand1），但在第一个外部循环完成后，它会断开：

标签[5]不在[index]
中

df['date'] = ''
df['totals'] = 0
months = ['Jan-17', 'Feb-17', 'Mar-17', 'Apr-17', 'May-17']

dropRowIndex = 0
nextDuplicateRowStartIndex = 0
totalRows = df.shape[0]

for i in range(0, totalRows):
    print('--------------')
    print(df)
    for col in df:
        if col in months:
            # Insert a row above 0th index with 0th row's values
            # Duplicate the row at this index for each month
            # Then move on to the next "row", which would be the latest index count
            df.loc[nextDuplicateRowStartIndex-1] = df.loc[nextDuplicateRowStartIndex].values
            df.loc[nextDuplicateRowStartIndex-1, 'date'] = col
            df.loc[nextDuplicateRowStartIndex-1, 'totals'] = df.loc[nextDuplicateRowStartIndex-1][col]

            df.index = df.index + 1
            df = df.sort_index()
            dropRowIndex += 1

    # Drop duplicated row by index
    df.drop(dropRowIndex, inplace=True)
    nextDuplicateRowStartIndex = dropRowIndex

# Remove months columns
for col in df:
    if col in months:
        df = df.drop(col, 1)

终端输出：

-------------- INITIAL DATA FRAME:
    brand  Jan-17  Feb-17  Mar-17  Apr-17  May-17 date  totals
0  brand1     222     333     444     555     666            0
1  brand2    7777    8888    9999    1010    1111            0
2  brand3   12121   13131   14141   15151   16161            0
-------------- DATA FRAME AFTER FIRST OUTER LOOP (ROW) ITERATION:
    brand  Jan-17  Feb-17  Mar-17  Apr-17  May-17    date  totals
0  brand1     222     333     444     555     666  May-17     666
1  brand1     222     333     444     555     666  Apr-17     555
2  brand1     222     333     444     555     666  Mar-17     444
3  brand1     222     333     444     555     666  Feb-17     333
4  brand1     222     333     444     555     666  Jan-17     222
6  brand2    7777    8888    9999    1010    1111               0
7  brand3   12121   13131   14141   15151   16161               0
Traceback (most recent call last):
  File "/Users/danielturcotte/Sites/project/env/lib/python3.6/site-packages/pandas/core/indexing.py", line 1506, in _has_valid_type
    error()
  File "/Users/danielturcotte/Sites/project/env/lib/python3.6/site-packages/pandas/core/indexing.py", line 1501, in error
    axis=self.obj._get_axis_name(axis)))
KeyError: 'the label [5] is not in the [index]'

ERROR

KeyError：＆＃39;标签[5]不在[index]＆＃39;

我有一个想法是因为我使用.loc[index]，其中index是一个整数，可能是.loc doesn't work with integers，但是.iloc[]。如果我做

df.iloc[nextDuplicateRowStartIndex-1] = df.iloc[nextDuplicateRowStartIndex].values

我收到错误：

ValueError：标签[10]未包含在轴
中

终端输出产生NaN s：

    brand  Jan-17  Feb-17  Mar-17  Apr-17  May-17    date  totals
0     NaN     NaN     NaN     NaN     NaN     NaN  May-17     NaN
1     NaN     NaN     NaN     NaN     NaN     NaN  Apr-17     NaN
2     NaN     NaN     NaN     NaN     NaN     NaN  Mar-17     NaN
3     NaN     NaN     NaN     NaN     NaN     NaN  Feb-17     NaN
4     NaN     NaN     NaN     NaN     NaN     NaN  Jan-17     NaN
6  brand2  7777.0  8888.0  9999.0  1010.0  1111.0             0.0
7     NaN     NaN     NaN     NaN     NaN     NaN  Apr-17     NaN

虽然我不相信这是问题，因为print(df.iloc[0])和print(df.loc[0])会产生相同的结果（即使我用整数访问loc[0]）

执行melt：

Answer 1

您可以使用melt。它允许您选择多个ID列和值列。在您的情况下，值列是除“品牌”之外的所有内容，因此我们可以忽略该参数。因此，您可以在一行中完成所有工作：

1. this.schedules = this.navigation.lineSelected.schedules;
2. this.schedules.filter (item => {
3. item> this.hourNow && item <this.hourFinish
4.});

打印：

import pandas as pd

df = pd.DataFrame({
    'brand': ['brand1', 'brand2', 'brand3'],
    'Jan-17': [22, 232, 324],
    'Feb-17': [333, 424, 999]
    # ...
})

rearranged = pd.melt(df, id_vars=['brand'], var_name='Date',
                     value_name='Total')

print(rearranged)

Answer 2

使用asongtoruin的数据和stack

df.set_index('brand').stack().reset_index(name='Total').rename(columns={'level_1':'Date'})
Out[1043]: 
    brand    Date  Total
0  brand1  Feb-17    333
1  brand1  Jan-17     22
2  brand2  Feb-17    424
3  brand2  Jan-17    232
4  brand3  Feb-17    999
5  brand3  Jan-17    324

＆＃34;标签[5]不在[索引]＆＃34;复制行

2 个答案: