How can I extract all the instances of a regular expression pattern in PySpark dataframe?

Question

I have a StringType() column in a PySpark dataframe. I want to extract all the instances of a regexp pattern from that string and put them into a new column of ArrayType(StringType())

Suppose the regexp pattern is [a-z]*([0-9]*)

Input df:
+-----------+
|stringValue|
+-----------+
|a1234bc123 |
|av1tb12h18 |
|abcd       | 
+-----------+

Output df:
+-----------+-------------------+
|stringValue|output             |
+-----------+-------------------+
|a1234bc123 |['1234', '123']    |
|av1tb12h18 |['1', '12', '18']  |
|abcd       |[]                 |
+-----------+-------------------+

Asked By: ARCrow

||

Source

Answer 1

Try use the split and array_remove from functions in spark:

create the test DataFrame

from pyspark.sql import functions as F
df = spark.createDataFrame([("a1234bc123",), ("av1tb12h18",), ("abcd",)],["stringValue"])
df.show()

The original DataFrame:

+-----------+
|stringValue|
+-----------+
| a1234bc123|
| av1tb12h18|
|       abcd|
+-----------+

Use split to separate the strings into numbers only

df = df.withColumn("mid", F.split('stringValue', r'[a-zA-Z]'))
df.show()

The output:

+-----------+-----------------+
|stringValue|              mid|
+-----------+-----------------+
| a1234bc123|  [, 1234, , 123]|
| av1tb12h18|[, , 1, , 12, 18]|
|       abcd|       [, , , , ]|
+-----------+-----------------+

Finally, use array_remove to remove non-number elements

df = df.withColumn("output", F.array_remove('mid', ''))
df.show()

The final output:

+-----------+-----------------+-----------+
|stringValue|              mid|     output|
+-----------+-----------------+-----------+
| a1234bc123|  [, 1234, , 123]|[1234, 123]|
| av1tb12h18|[, , 1, , 12, 18]|[1, 12, 18]|
|       abcd|       [, , , , ]|         []|
+-----------+-----------------+-----------+

Answered By: niuer

Answer 2

you can use the combination of regexp_replace and split api of function module

import pyspark.sql.types as t
import pyspark.sql.functions as f

l1 = [('anystring',),('a1234bc123',),('av1tb12h18',)]
df = spark.createDataFrame(l1).toDF('col')
df.show()
+----------+
|       col|
+----------+
| anystring|
|a1234bc123|
|av1tb12h18|
+----------+

Now use replace matching regex and then split by “,”. Here $1 refers to what is replaced value so it will be blank for matching regex.

e.g replace('anystring')
$0 = anystring
$1 = ""

dfl1 = df.withColumn('temp', f.split(f.regexp_replace("col", "[a-z]*([0-9]*)", "$1,"), ","))

dfl1.show()
+----------+---------------+
|       col|           temp|
+----------+---------------+
| anystring|         [, , ]|
|a1234bc123|[1234, 123, , ]|
|av1tb12h18|[1, 12, 18, , ]|
+----------+---------------+

Spark <2.4

Use UDF to replace empty values of array

def func_drop_from_array(arr):
    return [x for x in arr if x != '']

drop_from_array = f.udf(func_drop_from_array, t.ArrayType(t.StringType()))

dfl1.withColumn('final', drop_from_array('temp')).show()
+----------+---------------+-----------+
|       col|           temp|      final|
+----------+---------------+-----------+
| anystring|         [, , ]|         []|
|a1234bc123|[1234, 123, , ]|[1234, 123]|
|av1tb12h18|[1, 12, 18, , ]|[1, 12, 18]|
+----------+---------------+-----------+

Spark >=2.4

Use array_remove

dfl1.withColumn('final', f.array_remove('temp','')).show()

+----------+---------------+-----------+
|       col|           temp|      final|
+----------+---------------+-----------+
| anystring|         [, , ]|         []|
|a1234bc123|[1234, 123, , ]|[1234, 123]|
|av1tb12h18|[1, 12, 18, , ]|[1, 12, 18]|
+----------+---------------+-----------+

Answered By: SMaZ

Answer 3

In Spark 3.1+ regexp_extract_all is available.

regexp_extract_all(str, regexp[, idx]) – Extract all strings in the str that match the regexp expression and corresponding to the regex group index.

df = df.withColumn('output', F.expr(r"regexp_extract_all(stringValue, '[a-z]*(\d+)', 1)"))

df.show()
#+-----------+-----------+
#|stringValue|     output|
#+-----------+-----------+
#| a1234bc123|[1234, 123]|
#| av1tb12h18|[1, 12, 18]|
#|       abcd|         []|
#+-----------+-----------+

Answered By: ZygD

How can I extract all the instances of a regular expression pattern in PySpark dataframe?

Question:

Answers: