Odtnhj

I have a large table which is a union of several small selects as below:

create_sql = str("""
drop table if exists public.segments_table;
create table public.segments_table as

 select
 household_id,
 case
 when household_id in (select distinct household_id) from workmsmgr.ps_churn where drform = 'VOLUNTARY' and dt in (201804))
 then '1-3 Month Disco'
 else 'No Disco'
 end as disco_status,'201804' as churn_date,
 b.*
from
 public.churn_table b
 where drform = 'ACTIVE'
 and dt = 201803
 and tenure > 12

union all

select
 household_id,
 case
 when household_id in (select distinct household_id) from workmsmgr.ps_churn where drform = 'VOLUNTARY' and dt in (201805))
 then '1-3 Month Disco'
 else 'No Disco'
 end as disco_status,'201805' as churn_date,
 b.*
from
 public.churn_table b
 where drform = 'ACTIVE'
 and dt = 201803
 and tenure > 12

union all

 select
 household_id,

 case
 when household_id in (select distinct household_id) from workmsmgr.ps_churn where drform = 'VOLUNTARY' and dt in (201806))
 then '1-3 Month Disco'
 else 'No Disco'
 end as disco_status,'201806' as churn_date,
 b.*
from
 public.churn_table b
 where drform = 'ACTIVE'
 and dt = 201803
 and tenure > 12;
""")

#Now for each dt ,I pull the relevant records as below in python:

target_column = 'disco_status'
table_name = 'public.segments_table'

# Running for segments
query_params = 'month_date' : '201803',
 'churn_date' : '201804',
 'disco_target' : '1-3 Month Disco',
 'Y0' : 'No Disco'
 

sql_data_sample = str("""select * from 1
 where dt = %(month_date)s 
 and churn_date = %(churn_date)s
 and 0 IN (%(disco_target)s,'No Disco')
 ;""").format(target_column,table_name)

All these are part of a script called sql_data_prep.py which is imported in the main script:

from sql_data_prep import sql_data_sample,query_params,create_sql

#then a connection to the server is established and the query sql_data_sample is used to fetch various segments based on month_date and churn_date

database = 'Redshift Test'

print(datetime.datetime.strftime(datetime.datetime.now(), '%Y-%m-%d %H:%M:%S'))
try:
 cnxn = psycopg2.connect(database='xxxx',host='localhost',port='8880',user='dbo',password='xxxx')
 curs = cnxn.cursor()
 print ('*'*10 + "nConnected to %s!"%(database))
 print(datetime.datetime.strftime(datetime.datetime.now(), '%Y-%m-%d %H:%M:%S'))
 ##############################################################################
 #Parameterized Query: table_name and target_var to be also made as parameters.
 ##############################################################################

 curs.execute(create_sql)
 cnxn.commit()
 print('*'*10 + "nDone creating table with data")
 print(datetime.datetime.strftime(datetime.datetime.now(), '%Y-%m-%d %H:%M:%S'))

 df_data_sample = pd.read_sql(sql_data_sample,con = cnxn,params = query_params)

Now as you can see , I need to change the parameters thrice in query_params for three different segments :

1. Where dt = 201803 and churn_date = 201804
2. Where dt = 201803 and churn_date = 201805
3. Where dt = 201803 and churn_date = 201806

But I want to pass these as parameters and pull each segment sequentially in a for loop.
So my new query_params dict will look like:

query_params = 'month_date' : '201803',
 'churn_date' : ['201804','201805,'201806'],
 'disco_target' : '1-3 Month Disco',
 'Y0' : 'No Disco'
 

Once I connect to the database, I can:

for month in query_params['month_date]:
 for churn in query_params['churn_date']:
 sql_data_sample = str("""select * from 1
 where dt = %(month)s 
 and churn_date = %(churn)s
 and 0 IN (%(disco_target)s,'No Disco')
 ;""").format(target_column,table_name)

 # Some data processing functions:

This is being done to automate the process of creating tables from unions of several selects and then pull the segments sequentially, do some data processing
and move on to the next segment(which is defined by month_date and churn_date here.

I hope my question is clear and can someone please help me with this?