Comparison tpcds yql/pg (#4502)

Author: resetius

ydb/library/benchmarks/queries/tpcds/pg/q04.sql

@@ -111,7 +111,7 @@ union all
  order by t_s_secyear.customer_id
          ,t_s_secyear.customer_first_name
          ,t_s_secyear.customer_last_name
-         ,t_s_secyear.customer_email_address
+         ,t_s_secyear.customer_birth_country
 limit 100;
 
 -- end query 1 in stream 0 using template ../query_templates/query4.tpl

ydb/library/benchmarks/queries/tpcds/pg/q26.sql

@@ -11,7 +11,7 @@ select  i_item_id,
        cs_bill_cdemo_sk = cd_demo_sk and
        cs_promo_sk = p_promo_sk and
        cd_gender = 'F' and
-       cd_marital_status = 'W' and
+       cd_marital_status = 'S' and
        cd_education_status = 'College' and
        (p_channel_email = 'N' or p_channel_event = 'N') and
        d_year = 1998

ydb/library/benchmarks/queries/tpcds/pg/q70.sql

@@ -34,7 +34,7 @@ select
 ) as sub
  order by
    lochierarchy desc
-  ,case when lochierarchy = 0 then s_state end
+  ,case when lochierarchy = 0 then s_state else null end nulls first
   ,rank_within_parent
  limit 100;
 

ydb/library/benchmarks/queries/tpcds/yql/q15.sql

@@ -15,7 +15,7 @@ select  customer_address.ca_zip
      cross join {{date_dim}} as date_dim
  where cs_bill_customer_sk = c_customer_sk
  	and c_current_addr_sk = ca_address_sk
- 	and ( substring(cast(ca_zip as string),1,5) in ('85669', '86197','88274','83405','86475',
+ 	and ( substring(cast(ca_zip as string),0,5) in ('85669', '86197','88274','83405','86475',
                                    '85392', '85460', '80348', '81792')
  	      or ca_state in ('CA','WA','GA')
  	      or cs_sales_price > 500)

ydb/library/benchmarks/queries/tpcds/yql/q19.sql

@@ -17,7 +17,7 @@ select  item.i_brand_id brand_id, item.i_brand brand, item.i_manufact_id, item.i
    and d_year=1998
    and ss_customer_sk = c_customer_sk
    and c_current_addr_sk = ca_address_sk
-   and substring(cast(ca_zip as string),1,5) <> substring(cast(s_zip as string),1,5)
+   and substring(cast(ca_zip as string),0,5) <> substring(cast(s_zip as string),0,5)
    and ss_store_sk = s_store_sk
  group by item.i_brand
       ,item.i_brand_id

ydb/library/benchmarks/queries/tpcds/yql/q23.sql

@@ -2,7 +2,7 @@
 
 -- NB: Subquerys
 $blabla = (
-    select substring(cast(item.i_item_desc as string),1,30) itemdesc,item.i_item_sk item_sk,date_dim.d_date solddate
+    select substring(cast(item.i_item_desc as string),0,30) itemdesc,item.i_item_sk item_sk,date_dim.d_date solddate
   from {{store_sales}} as store_sales
       cross join {{date_dim}} as date_dim
       cross join {{item}} as item

ydb/library/benchmarks/queries/tpcds/yql/q45.sql

@@ -11,7 +11,7 @@ select  customer_address.ca_zip, customer_address.ca_county, sum(ws_sales_price)
  where ws_bill_customer_sk = c_customer_sk
  	and c_current_addr_sk = ca_address_sk
  	and ws_item_sk = i_item_sk
- 	and ( substring(cast(ca_zip as string),1,5) in ('85669', '86197','88274','83405','86475', '85392', '85460', '80348', '81792')
+ 	and ( substring(cast(ca_zip as string),0,5) in ('85669', '86197','88274','83405','86475', '85392', '85460', '80348', '81792')
  	      or
  	      i_item_id in (select i_item_id
                              from {{item}} as item

ydb/library/benchmarks/queries/tpcds/yql/q62.sql

@@ -27,7 +27,7 @@ and ws_warehouse_sk   = w_warehouse_sk
 and ws_ship_mode_sk   = sm_ship_mode_sk
 and ws_web_site_sk    = web_site_sk
 group by
-   substring(cast(w_warehouse_name as string),1,20) as bla
+   substring(cast(w_warehouse_name as string),0,20) as bla
   ,ship_mode.sm_type
   ,web_site.web_name
 order by bla

ydb/library/benchmarks/queries/tpcds/yql/q79.sql

@@ -3,7 +3,7 @@
 -- NB: Subquerys
 -- start query 1 in stream 0 using template query79.tpl and seed 2031708268
 select
-  c_last_name,c_first_name,substring(cast(s_city as string),1,30) bla,ss_ticket_number,amt,profit
+  c_last_name,c_first_name,substring(cast(s_city as string),0,30) bla,ss_ticket_number,amt,profit
   from
    (select store_sales.ss_ticket_number ss_ticket_number
           ,store_sales.ss_customer_sk ss_customer_sk

ydb/library/benchmarks/queries/tpcds/yql/q8.sql

@@ -1,9 +1,9 @@
 {% include 'header.sql.jinja' %}
 
 -- NB: Subquerys
-$bla1 = (      SELECT substring(cast(ca_zip as string),1,5) AS ca_zip
+$bla1 = (      SELECT substring(cast(ca_zip as string),0,5) AS ca_zip
       FROM {{customer_address}}
-      WHERE substring(cast(ca_zip as string),1,5) IN (
+      WHERE substring(cast(ca_zip as string),0,5) IN (
                           '47602','16704','35863','28577','83910','36201',
                           '58412','48162','28055','41419','80332',
                           '38607','77817','24891','16226','18410',
@@ -85,7 +85,7 @@ $bla1 = (      SELECT substring(cast(ca_zip as string),1,5) AS ca_zip
                           '28239','58032','18884','16791','21343',
                           '97462','18569','75660','15475'));
 $bla2 = (select A1.ca_zip as ca_zip
-      from (SELECT substring(cast(customer_address.ca_zip as string),1,5) ca_zip,count(*) cnt
+      from (SELECT substring(cast(customer_address.ca_zip as string),0,5) ca_zip,count(*) cnt
             FROM {{customer_address}} as customer_address
             cross join {{customer}} as customer
             WHERE ca_address_sk = c_current_addr_sk and
@@ -107,7 +107,7 @@ select  store.s_store_name
  where ss_store_sk = s_store_sk
   and ss_sold_date_sk = d_date_sk
   and d_qoy = 2 and d_year = 1998
-  and (substring(cast(s_zip as string),1,2) = substring(V1.ca_zip,1,2))
+  and (substring(cast(s_zip as string),0,2) = substring(V1.ca_zip,0,2))
  group by store.s_store_name
  order by store.s_store_name
  limit 100;