如何在 Apache Spark 中的 Dataframe 上运行 Python 中的正则表达式

Question

I'm trying run a Regex in Python on a dataframe in Apache Spark.

The df is

The regex is as follows:

import re
m = re.search("[Pp]ython", df)
print(m)

I'm getting the following error message:

TypeError: expected string or bytes-like object

The following will work

import re m = re.search("[Pp]ython", 'Python python') print(m)

But I would like the regex to work on a dataframe

Answer 1

You can use regexp_extract :

from pyspark.sql import functions as F

data = [["Python"],["python"], ["Scala"], ["PYTHON"]]
schema= ["language"]

df = spark.createDataFrame(data, schema)

df = df.withColumn("extracted", F.regexp_extract("language", "[Pp]ython", 0))

Result:

+--------+---------+
|language|extracted|
+--------+---------+
|  Python|   Python|
|  python|   python|
|   Scala|         |
|  PYTHON|         |
+--------+---------+

The definition for re.search is

re.search(pattern, string, flags=0)

The second parameter being a string, this function cannot work with Spark dataframes. However (at least most) patterns that work with re.search will also work for regexp_extract . So testing the patterns with re.search first might be a way.