创建新的 pd 数据框列，该列根据起始数据的日和周给出日期

Question

我有一个 pandas 数据框，它有两列，第一列是 'Week Starting'，另一列是 'Day'。我想创建一个新列，使用其他两列的数据来提供完整日期。例如，在下面的 table 中，新列的第一个条目应该是 5/04/2021，第二个条目应该是 6/04/2021。

Week Starting	Day
5/04/2021	Monday
5/04/2021	Tuesday
5/04/2021	Wednesday

我尝试了以下解决方案，但出现错误

g['Week Starting'] = pd.to_datetime(g['Week Starting'])

conditions = [ (g['Day'] == 'Monday'), (g['Day'] == 'Tuesay'), (g['Day'] == 
                'Wednesday')]

values = [g['Week Starting'],(g['Week Starting'] + timedelta(days=1)), 
          (g['Week Starting'] + timedelta(days=2))]

g['Date'] = np.select(conditions, values)

错误：

DType 和没有通用的 DType。例如，除非 dtype 为 object.

，否则它们不能存储在单个数组中

谢谢。

Answer 1

我认为这是最简单的解决方案：

df = pd.DataFrame({"week_starting":["04/05/2021","04/05/2021","04/05/2021"],
                    "day":["Monday","Tuesday","Wednesday"]})

df['week_starting'] = pd.to_datetime(df['week_starting'])

conditions = {"Monday":0,"Tuesday":1,"Wednesday":2}

df["date"] = df.apply(lambda x:x['week_starting']+pd.Timedelta(conditions[x["day"]],"day"),axis=1)

您使用 apply 方法将时间增量添加到每个日期。

希望有用！

Answer 2

使用to_timedelta with mapping values by Series.map:

df['week_starting'] = pd.to_datetime(df['week_starting'])

d = {"Monday":0,"Tuesday":1,"Wednesday":2,
     "Thursday":3,"Friday":4,"Saturday":5, 'Sunday':6}

df["date"] = df['week_starting'] + pd.to_timedelta(df["day"].map(d),"day")
print (df)
  week_starting        day       date
0    2021-04-05     Monday 2021-04-05
1    2021-04-05    Tuesday 2021-04-06
2    2021-04-05  Wednesday 2021-04-07

创建新的 pd 数据框列，该列根据起始数据的日和周给出日期

Create new pd dataframe column that gives a date based on day and week starting data

python

datetime

dataframe

pandas

data-science