Pyspark将年和周数转换为周开始日期和周结束日期

2024-09-24 10:19:53 发布

您现在位置:Python中文网/ 问答频道 /正文

我正在尝试转换year&;使用Pyspark数据框将周数设置为周开始日期和周结束日期

+---------+
|year_week|
+---------+
| 2019-W51|
| 2019-W52|
| 2020-W01|
| 2020-W02|
| 2020-W03|
| 2020-W04|
| 2020-W05|
| 2020-W06|
| 2020-W07|
+---------+

当我尝试应用下面的代码时,我得到的错误是column对象不可编辑

df = df.withColumn('week_start_date', df.year_week.apply(lambda x: datetime.datetime.strptime(d + '-1', "%Y-W%W-%w")))

错误

TypeError: 'Column' object is not callable

预期结果是:

+---------+----------+----------+
|year_week|week_start|  week_end|
+---------+----------+----------+
| 2019-W51|2019-12-16|2019-12-22|
| 2019-W52|2019-12-23|2019-12-29|
| 2020-W01|2019-12-30|2020-01-05|
| 2020-W02|2020-01-06|2020-01-12|
| 2020-W03|2020-01-13|2020-01-19|
| 2020-W04|2020-01-20|2020-01-26|
| 2020-W05|2020-01-27|2020-02-02|
| 2020-W06|2020-02-03|2020-02-09|
| 2020-W07|2020-02-10|2020-02-16|
+---------+----------+----------+


Tags: dfdatetime错误yearstartweekw03w02
2条回答

试试这个-

 // week starting from monday, concat "-1", for tuesday "-2" etc.
    val p = df2.withColumn("week_start", to_date(concat($"year_week", lit("-1")), "YYYY-'W'ww-u"))
      .withColumn("week_end", next_day($"week_start", "SUN"))
    p.show(false)
    p.printSchema()

    /**
      * +    -+     +     +
      * |year_week|week_start|week_end  |
      * +    -+     +     +
      * |2019-W51 |2019-12-16|2019-12-22|
      * |2019-W52 |2019-12-23|2019-12-29|
      * |2020-W01 |2019-12-30|2020-01-05|
      * |2020-W02 |2020-01-06|2020-01-12|
      * |2020-W03 |2020-01-13|2020-01-19|
      * |2020-W04 |2020-01-20|2020-01-26|
      * |2020-W05 |2020-01-27|2020-02-02|
      * |2020-W06 |2020-02-03|2020-02-09|
      * |2020-W07 |2020-02-10|2020-02-16|
      * +    -+     +     +
      *
      * root
      * |  year_week: string (nullable = true)
      * |  week_start: date (nullable = true)
      * |  week_end: date (nullable = true)
      */

根据Someshwar的反应,只需对scala中的python进行细微的更改

df1 = df.withColumn("week_start", F.to_date(F.concat(F.col("year_week"), F.lit("-1")), "YYYY-'W'ww-u")).withColumn("week_end", F.next_day(F.col("week_start"), "SUN"))

相关问题 更多 >