dataframe - How does for loop impact performance of spark code

Question

I have two piece of codes below having same logic, curious to know which one is better of two and why?

1.

char_list = [('\\', '\\\\'), ('
', '\\n'), (''', '\\'')]
col_names = df.schema.names
df.select( *[func.regexp_replace(col_name, char_set[0], char_set[1]) for char_set in char_list for col_name in col_names])

char_list = [('\\', '\\\\'), ('
', '\\n'), (''', '\\'')]
col_names = df.schema.names
for char_set in char_list:
    for col_name in col_names:
        df = df.withColumn(col_name, func.regexp_replace(col_name, char_set[0], char_set[1]))

Answer 1

The logic of the two codes are not the same. The second code should be what you wanted. In the first code you selected duplicated columns because select does not overwrite columns, but withColumn does.

import pyspark.sql.functions as func

char_list = [('\\', '\\\\'), ('
', '\\n'), (''', '\\'')]
col_names = df.schema.names

df = spark.createDataFrame([['1','2']])
print(len(df.select( *[func.regexp_replace(col_name, char_set[0], char_set[1]) for char_set in char_list for col_name in col_names]).columns))
# gives 6

df = spark.createDataFrame([['1','2']])
for char_set in char_list:
    for col_name in col_names:
        df = df.withColumn(col_name, func.regexp_replace(col_name, char_set[0], char_set[1]))

print(len(df.columns))
# gives 2