Pyspark Scenarios 11 : how to handle double delimiter or multi delimiters in pyspark

[@sravankumar1767]

Nice explanation 👌 👍 👏

[@rajeshk1276]

Very Well explained.. Loved it

[@prabhakaranvelusamy]

Excellent explanation!

[@TRRaveendra]

Thank you 👍

[@gobinathmuralitharan1997]

Clear explanation 👍👏thank you 🙂

[@TRRaveendra]

Thank You Gobinath

[@pokemongatcha122]

Hi Ravi, I'm trying to do split by delimiter of a column with each cell having different no. of commas. Can you write a code to split columns with each occurance of comma? E.g. if row 1 has 4 commas it generates 4 columns but row 2 has 10 commas so it further generates another 6 columns.

[@tanushreenagar3116]

Nice

[@gobinathmuralitharan1997]

Subscribed 🔔

[@TRRaveendra]

Thank You 😍

[@vikrammore-y4t]

spark 3.X supports multi delimiter like .option("delimiter","[||]")

[@penchalaiahnarakatla9396]

Hi, good video, one clarification, while writing dataframe output to csv leading zeros are missing.. How to handle this secanioro. If possible make a video on this.

[@TRRaveendra]

Thank you 👍

[@penchalaiahnarakatla9396]

Hope next video will be this.

[@fakrullahimran]

@@penchalaiahnarakatla9396 Try to include option(“quoteAll”,True) and check once

[@penchalaiahnarakatla9396]

@@fakrullahimran Thanks. I will try and will update you..

[@udaynayak4788]

Hi Ravi, i do have .txt file which multiple space delimiter, e.g accountID Acctnbm acctadd branch and likewise can you please suggest the approach here almost i have 76 columns with multiple consecutive delimiter.

[@V-Barah]

this is looks simple in example but in real time we can't do each with column if there are 200-300 columns.
is there any other way?

[@DanishAnsari-hw7so]

# Case 1. when no of columns is known
col = 4
i = 0
while i < col:
df_multi = df_multi.withColumn("sub" + str(i), df_multi["marks_split"][i])
i += 1
df_1 = df_multi.drop("marks").drop("marks_split")
display(df_1)

[@DanishAnsari-hw7so]

# Case 2. when no of columns is not known known
from pyspark.sql.functions import max
df_multi = df_multi.withColumn('marks_size', size('marks_split'))
max_size = df_multi.select(max('marks_size')).collect()[0][0]
j = 0
while j < max_size:
df_multi = df_multi.withColumn("subject" + str(j), df_multi["marks_split"][j])
j += 1
df_2 = df_multi.drop("marks").drop("marks_split").drop('marks_size')
display(df_2)

[@snagendra5415]

Could you explain spark small files problem using pyspark?
Thank you in advance

[@TRRaveendra]

sure i will do video on small files problem.

[@snagendra5415]

@@TRRaveendra thank you for your reply, and waiting for the video 🤩

[@fratkalkan7850]

perfection

[@JustForFun-oy8fu]

Hi Ravi, thanks I have one doubt: how
can we generalize the above logic.....like if we have large number of columns after splitting the data like then it's obvious we can't do it manually.
What could be our approach in that case?
Thanks,
Anonymous

[@DanishAnsari-hw7so]

# Case 1. when no of columns is known
col = 4
i = 0
while i < col:
df_multi = df_multi.withColumn("sub" + str(i), df_multi["marks_split"][i])
i += 1
df_1 = df_multi.drop("marks").drop("marks_split")
display(df_1)

[@DanishAnsari-hw7so]

# Case 2. when no of columns is not known known
from pyspark.sql.functions import max
df_multi = df_multi.withColumn('marks_size', size('marks_split'))
max_size = df_multi.select(max('marks_size')).collect()[0][0]
j = 0
while j < max_size:
df_multi = df_multi.withColumn("subject" + str(j), df_multi["marks_split"][j])
j += 1
df_2 = df_multi.drop("marks").drop("marks_split").drop('marks_size')
display(df_2)

[@NaveenKumar-kb2fm]

very well explained , i have a scenario with schema (id,name,age,technology) and data in single row like (1001|Ram|28|Java|1002|Raj|24|Database|1004|Jam|28|DotNet|1005|Kesh|25|Java) coming in a single csv file.
now can we make it into multiple rows as per schema as a single table like below
id,name,age,technology
1001|Ram|28|Java
1002|Raj|24|Database
1004|Jam|28|DotNet
1005|Kesh|25|Java

[@mohitmotwani9256]

This data needs to be deived in multiple lines.

[@dinsan4044]

Hi ,
Could you please create a video to combine below 3 csv data files into one data frame dynamically
File name: Class_01.csv
StudentID Student Name Gender Subject B Subject C Subject D
1 Balbinder Male 91 56 65
2 Sushma Female 90 60 70
3 Simon Male 75 67 89
4 Banita Female 52 65 73
5 Anita Female 78 92 57
File name: Class_02.csv
StudentID Student Name Gender Subject A Subject B Subject C Subject E
1 Richard Male 50 55 64 66
2 Sam Male 44 67 84 72
3 Rohan Male 67 54 75 96
4 Reshma Female 64 83 46 78
5 Kamal Male 78 89 91 90
File name: Class_03.csv
StudentID Student Name Gender Subject A Subject D Subject E
1 Mohan Male 70 39 45
2 Sohan Male 56 73 80
3 shyam Male 60 50 55
4 Radha Female 75 80 72
5 Kirthi Female 60 50 55