CODEWITHSUNDEEP
helperdelete-rowcpu-wordpyspark
andrea perez
andrea perez

Reputation: 43

remove specific words into a dataframe with pyspark

I have a DataFrame 

+------+--------------------+-----------------+----
|   id| titulo       |tipo      | formacion       |
+------+--------------------+-----------------+----
|32084|A             | Material | VION00001 TRADE |
|32350|B             | Curso    | CUS11222  LEADER|
|32362|C             | Curso    | ITIN9876  EVALUA|   
|32347|D             | Curso    | CUMPLI VION1234 |      
|32036|E             | Curso    | EVAN1111  INFORM|

I need, that into formacion column remove the characters that start with VION|CUS|ITIN|VION|EVAN so Dataframe looks like

+------+--------------------+-----------------+----
|   id| titulo       |tipo      | formacion       |
+------+--------------------+-----------------+----
|32084|A             | Material |  TRADE          |
|32350|B             | Curso    |  LEADER         |
|32362|C             | Curso    |  EVALUA         |   
|32347|D             | Curso    |  CUMPLI         |      
|32036|E             | Curso    |  INFORM         |  
+------+--------------------+-----------------+----

Thank you for your help

Upvotes: 3

Views: 311

Answers (2)

notNull
notNull

Reputation: 31510

Use split function to split the column by space then get the last element of array.

  • from Spark2.4+ use element_at function
  • for Spark < 2.4 use reverse(split(array))[0]
#using element_at
df.withColumn("formacion",element_at(split(col("formacion"),"\\s"),-1)).show() 

#or using array_index
df.withColumn("formacion",split(col("formacion"),"\\s")[1]).show()

#split reverse and get first index value
df.withColumn("formacion",reverse(split(col("formacion"),"\\s"))[0]).show()

#+-----+--------------+----------+-------------+
#|   id|titulo        |tipo      | formacion   |
#+------+--------------------+-----------------+
#|32084|A             | Material |  TRADE      |
#|32350|B             | Curso    |  LEADER     |
#|32362|C             | Curso    |  EVALUA     |   
#|32347|D             | Curso    |  CUMPLI     |      
#|32036|E             | Curso    |  INFORM     |  
#+-----+--------------+----------+-------------+

Upvotes: 2

andrea perez
andrea perez

Reputation: 43

Sorry guys , this is original Column from DataFrame

formacion = [ VION00001 TRADE, CUS11222 LEADER,ITIN9876 EVALUA ,VION1234 CUMPLI,EVAN11 FR]

This is expected

formacion = [ TRADE, LEADER,EVALUA ,CUMPLI, FR]

Upvotes: 0

Related Questions