Regex 如何在scala（spark）中提取特定字符串后的值？

我有一个带有列的数据帧：

df=

我需要得到：

itemType                   count
it_shampoo                  5
it_books                    5
it_mm                       5
it_mm                       5
it_books                    5
it_books                    5

如何提取将

it\u books it\u books

，

{=it\u books}it\u books

替换为

it\u books

。项目类型将始终跟随

it

尝试regex，

^.*（it[\w]+）.$

到项目类型，并替换为第一个捕获的组

$1

---

下面的正则表达式也适用

scala> val df = Seq(("it_shampoo",5),
     | ("it_books",5),
     | ("it_mm",5),
     | ("{it_mm}",5),
     | ("it_books it_books",5),
     | ("{=it_books} it_books",5)).toDF("itemType","count")
df: org.apache.spark.sql.DataFrame = [itemType: string, count: int]

scala> df.select( regexp_replace('itemtype,""".*\b(\S+)\b(.*)$""", "$1").as("replaced"),'count).show
+----------+-----+
|  replaced|count|
+----------+-----+
|it_shampoo|    5|
|  it_books|    5|
|     it_mm|    5|
|     it_mm|    5|
|  it_books|    5|
|  it_books|    5|
+----------+-----+


scala>

---

你的正则表达式为我工作。我使用的代码是

val df2=df.withColumn（“itemType”，regexp\u extract（$“itemType”，即“^.*”）。（it.[\w]+）.$”，1））

scala> val df = Seq(("it_shampoo",5),
     | ("it_books",5),
     | ("it_mm",5),
     | ("{it_mm}",5),
     | ("it_books it_books",5),
     | ("{=it_books} it_books",5)).toDF("itemType","count")
df: org.apache.spark.sql.DataFrame = [itemType: string, count: int]

scala> df.select( regexp_replace('itemtype,""".*\b(\S+)\b(.*)$""", "$1").as("replaced"),'count).show
+----------+-----+
|  replaced|count|
+----------+-----+
|it_shampoo|    5|
|  it_books|    5|
|     it_mm|    5|
|     it_mm|    5|
|  it_books|    5|
|  it_books|    5|
+----------+-----+


scala>