converting Informatica transformations to Pyspark

Question

I am trying to convert informatica transformation to pyspark transformation, but I am stuck in replacing char in the code shown below:

"DECODE(TRUE,
ISNULL(v_check_neg_**) OR v_check_neg_** = '', 
i_default,
NOT IS_NUMBER(v_check_neg_** , 
i_default,
REPLACECHR(0,v_check_neg_**, '+-0123456789.' ,'')<>'', 
i_default,
TO_DECIMAL(v_check_neg_**,5))


v_check_neg_** = IIF(INSTR(i_string_**,'-')!=0,'-'||SUBSTR(i_string_**,1,INSTR(i_string_**,'-')-1),i_string_**)"

This is what I tried:

def is_digit(value):
    if value:
        return value.isdigit()
    else:
        return False

is_digit_udf = udf(is_digit, BooleanType())

df_informatica=df_informatica.withColumn(column_name,when((isnull(col(column_name)) |(col(column_name==' ')),i_default).when(is_digit_udf(col(column_name)),i_default)

df_informatica=df_informatica.withColumn

Please help me convert informatica to pyspark transformation.

Answer 1

i cant see whole statement but your decode logic is - if (v_check_neg_** is null or v_check_neg_ ='' or v_check_neg_ is not number or v_check_neg_** has anything other than numbers) then i_default else TO_DECIMAL(v_check_neg_**,5)

Use python to check above cases and you should be good to go. Like you can use string.isnumeric() to check positive number. And use try except to check -ve, decimal etc. Example to check negative number-

def check_negative(s):
    try:
        f = float(s)
        if (f < 0):
            return True
        # Otherwise return false
        return False
    except ValueError:
        return False