Modify verification queries

Author: pratyakshsharma

benchmarks/tpc-ds/queries/refresh/verify_LF_CR.sql

@@ -1,84 +1,41 @@
 -- ==============================================
---    Fetch two rows randomly from crv:
+-- 144662
+-- 08 39 4f 4a 5b b6 95 4e
 -- ==============================================
-with crv_random as (
-    select
-        cr_return_date_sk,
-        cr_return_time_sk,
-        cr_item_sk,
-        cr_refunded_customer_sk,
-        cr_refunded_cdemo_sk,
-        cr_refunded_hdemo_sk,
-        cr_refunded_addr_sk,
-        cr_returning_customer_sk,
-        cr_returning_cdemo_sk,
-        cr_returning_hdemo_sk,
-        cr_returning_addr_sk,
-        cr_call_center_sk,
-        cr_catalog_page_sk,
-        cr_ship_mode_sk,
-        cr_warehouse_sk,
-        cr_reason_sk,
-        cr_order_number,
-        cr_return_quantity,
-        cr_return_amount,
-        cr_return_tax,
-        cr_return_amt_inc_tax,
-        cr_fee,
-        cr_return_ship_cost,
-        cr_refunded_cash,
-        cr_reversed_charge,
-        cr_merchant_credit,
-        cr_net_loss
-    from crv
-    where cr_return_time_sk >= (select floor( max(cr_return_time_sk) * rand()) from crv )
-    order by cr_return_time_sk limit 2)
--- ========================================================
---    Verify the row can be selected from catalog_returns:
--- ========================================================
-select
-    cr.cr_returned_date_sk,
-    cr.cr_returned_time_sk,
-    cr.cr_item_sk,
-    cr.cr_refunded_customer_sk,
-    cr.cr_refunded_cdemo_sk,
-    cr.cr_refunded_hdemo_sk,
-    cr.cr_refunded_addr_sk,
-    cr.cr_returning_customer_sk,
-    cr.cr_returning_cdemo_sk,
-    cr.cr_returning_hdemo_sk,
-    cr.cr_returning_addr_sk,
-    cr.cr_call_center_sk,
-    cr.cr_catalog_page_sk,
-    cr.cr_ship_mode_sk,
-    cr.cr_warehouse_sk,
-    cr.cr_reason_sk,
-    cr.cr_order_number,
-    cr.cr_return_quantity,
-    cr.cr_return_amount,
-    cr.cr_return_tax,
-    cr.cr_return_amt_inc_tax,
-    cr.cr_fee,
-    cr.cr_return_ship_cost,
-    cr.cr_refunded_cash,
-    cr.cr_reversed_charge,
-    cr.cr_store_credit,
-    cr.cr_net_loss
-from catalog_returns cr, crv_random
-where
-        cr.cr_returned_date_sk = crv_random.cr_return_date_sk and
-        cr.cr_returned_time_sk = crv_random.cr_return_time_sk and
-        cr.cr_item_sk = crv_random.cr_item_sk and
-        cr.cr_refunded_customer_sk = crv_random.cr_refunded_customer_sk and
-        cr.cr_refunded_cdemo_sk = crv_random.cr_refunded_cdemo_sk and
-        cr.cr_refunded_hdemo_sk = crv_random.cr_refunded_hdemo_sk and
-        cr.cr_refunded_addr_sk = crv_random.cr_refunded_addr_sk and
-        cr.cr_returning_customer_sk = crv_random.cr_returning_customer_sk and
-        cr.cr_returning_cdemo_sk = crv_random.cr_returning_cdemo_sk and
-        cr.cr_returning_hdemo_sk  = crv_random.cr_returning_hdemo_sk and
-        cr.cr_returning_addr_sk = crv_random.cr_returning_addr_sk and
-        cr.cr_call_center_sk = crv_random.cr_call_center_sk and
-        cr.cr_catalog_page_sk = crv_random.cr_catalog_page_sk and
-        cr.cr_ship_mode_sk = crv_random.cr_ship_mode_sk and
-        cr.cr_warehouse_sk = crv_random.cr_warehouse_sk and
-        cr.cr_reason_sk = crv_random.cr_reason_sk;
+select count(*) from catalog_returns;
+
+with column_checksums as (
+    select array[
+               checksum(cr_returned_date_sk),
+           checksum(cr_returned_time_sk),
+           checksum(cr_item_sk),
+           checksum(cr_refunded_customer_sk),
+           checksum(cr_refunded_cdemo_sk),
+           checksum(cr_refunded_hdemo_sk),
+           checksum(cr_refunded_addr_sk),
+           checksum(cr_returning_customer_sk),
+           checksum(cr_returning_cdemo_sk),
+           checksum(cr_returning_hdemo_sk),
+           checksum(cr_returning_addr_sk),
+           checksum(cr_call_center_sk),
+           checksum(cr_catalog_page_sk),
+           checksum(cr_ship_mode_sk),
+           checksum(cr_warehouse_sk),
+           checksum(cr_reason_sk),
+           checksum(cr_order_number),
+           checksum(cr_return_quantity),
+           checksum(cr_return_amount),
+           checksum(cr_return_tax),
+           checksum(cr_return_amt_inc_tax),
+           checksum(cr_fee),
+           checksum(cr_return_ship_cost),
+           checksum(cr_refunded_cash),
+           checksum(cr_reversed_charge),
+           checksum(cr_store_credit),
+           checksum(cr_net_loss)
+    ] checksums
+from catalog_returns
+    )
+select checksum(cs) as table_checksum
+from column_checksums
+         cross join unnest(column_checksums.checksums) as x(cs);

benchmarks/tpc-ds/queries/refresh/verify_LF_CS.sql

@@ -1,101 +1,48 @@
 -- ==============================================
---    Fetch two rows randomly from csv:
+-- 1447686
+-- fd 54 dc 73 48 5d da 70
 -- ==============================================
-with csc_random as
-         (select
-              cs_sold_date_sk,
-              cs_sold_time_sk,
-              cs_ship_date_sk,
-              cs_bill_customer_sk,
-              cs_bill_cdemo_sk,
-              cs_bill_hdemo_sk,
-              cs_bill_addr_sk,
-              cs_ship_customer_sk,
-              cs_ship_cdemo_sk,
-              cs_ship_hdemo_sk,
-              cs_ship_addr_sk,
-              cs_call_center_sk,
-              cs_catalog_page_sk,
-              cs_ship_mode_sk,
-              cs_warehouse_sk,
-              cs_item_sk,
-              cs_promo_sk,
-              cs_order_number,
-              cs_quantity,
-              cs_wholesale_cost,
-              cs_list_price,
-              cs_sales_price,
-              cs_ext_discount_amt,
-              cs_ext_sales_price,
-              cs_ext_wholesale_cost,
-              cs_ext_list_price,
-              cs_ext_tax,
-              cs_coupon_amt,
-              cs_ext_ship_cost,
-              cs_net_paid,
-              cs_net_paid_inc_tax,
-              cs_net_paid_inc_ship,
-              cs_net_paid_inc_ship_tax,
-              cs_net_profit
-          from csv
-          where cs_sold_time_sk >= (select floor( max(cs_sold_time_sk) * rand()) from csv )
-          order by cs_sold_time_sk limit 2)
+select count(*) from catalog_sales;
 
--- ========================================================
---    Verify the row can be selected from catalog_sales:
--- ========================================================
-
-select
-    catalog_sales.cs_sold_date_sk,
-    catalog_sales.cs_sold_time_sk,
-    catalog_sales.cs_ship_date_sk,
-    catalog_sales.cs_bill_customer_sk,
-    catalog_sales.cs_bill_cdemo_sk,
-    catalog_sales.cs_bill_hdemo_sk,
-    catalog_sales.cs_bill_addr_sk,
-    catalog_sales.cs_ship_customer_sk,
-    catalog_sales.cs_ship_cdemo_sk,
-    catalog_sales.cs_ship_hdemo_sk,
-    catalog_sales.cs_ship_addr_sk,
-    catalog_sales.cs_call_center_sk,
-    catalog_sales.cs_catalog_page_sk,
-    catalog_sales.cs_ship_mode_sk,
-    catalog_sales.cs_warehouse_sk,
-    catalog_sales.cs_item_sk,
-    catalog_sales.cs_promo_sk,
-    catalog_sales.cs_order_number,
-    catalog_sales.cs_quantity,
-    catalog_sales.cs_wholesale_cost,
-    catalog_sales.cs_list_price,
-    catalog_sales.cs_sales_price,
-    catalog_sales.cs_ext_discount_amt,
-    catalog_sales.cs_ext_sales_price,
-    catalog_sales.cs_ext_wholesale_cost,
-    catalog_sales.cs_ext_list_price,
-    catalog_sales.cs_ext_tax,
-    catalog_sales.cs_coupon_amt,
-    catalog_sales.cs_ext_ship_cost,
-    catalog_sales.cs_net_paid,
-    catalog_sales.cs_net_paid_inc_tax,
-    catalog_sales.cs_net_paid_inc_ship,
-    catalog_sales.cs_net_paid_inc_ship_tax,
-    catalog_sales.cs_net_profit
-from catalog_sales, csc_random
-where
-        catalog_sales.cs_sold_date_sk = csc_random.cs_sold_date_sk and
-        catalog_sales.cs_sold_time_sk = csc_random.cs_sold_time_sk and
-        catalog_sales.cs_ship_date_sk = csc_random.cs_ship_date_sk and
-        catalog_sales.cs_bill_customer_sk = csc_random.cs_bill_customer_sk and
-        catalog_sales.cs_bill_cdemo_sk = csc_random.cs_bill_cdemo_sk and
-        catalog_sales.cs_bill_hdemo_sk = csc_random.cs_bill_hdemo_sk and
-        catalog_sales.cs_bill_addr_sk = csc_random.cs_bill_addr_sk and
-        catalog_sales.cs_ship_customer_sk = csc_random.cs_ship_customer_sk and
-        catalog_sales.cs_ship_cdemo_sk = csc_random.cs_ship_cdemo_sk and
-        catalog_sales.cs_ship_hdemo_sk = csc_random.cs_ship_hdemo_sk and
-        catalog_sales.cs_ship_addr_sk = csc_random.cs_ship_addr_sk and
-        catalog_sales.cs_call_center_sk = csc_random.cs_call_center_sk and
-        catalog_sales.cs_catalog_page_sk = csc_random.cs_catalog_page_sk and
-        catalog_sales.cs_ship_mode_sk = csc_random.cs_ship_mode_sk and
-        catalog_sales.cs_warehouse_sk = csc_random.cs_warehouse_sk and
-        catalog_sales.cs_item_sk = csc_random.cs_item_sk and
-        catalog_sales.cs_promo_sk = csc_random.cs_promo_sk;
+with column_checksums as (
+    select array[
+               checksum(cs_sold_date_sk),
+           checksum(cs_sold_time_sk),
+           checksum(cs_ship_date_sk),
+           checksum(cs_bill_customer_sk),
+           checksum(cs_bill_cdemo_sk),
+           checksum(cs_bill_hdemo_sk),
+           checksum(cs_bill_addr_sk),
+           checksum(cs_ship_customer_sk),
+           checksum(cs_ship_cdemo_sk),
+           checksum(cs_ship_hdemo_sk),
+           checksum(cs_ship_addr_sk),
+           checksum(cs_call_center_sk),
+           checksum(cs_catalog_page_sk),
+           checksum(cs_ship_mode_sk),
+           checksum(cs_warehouse_sk),
+           checksum(cs_item_sk),
+           checksum(cs_promo_sk),
+           checksum(cs_order_number),
+           checksum(cs_quantity),
+           checksum(cs_wholesale_cost),
+           checksum(cs_list_price),
+           checksum(cs_sales_price),
+           checksum(cs_ext_discount_amt),
+           checksum(cs_ext_sales_price),
+           checksum(cs_ext_wholesale_cost),
+           checksum(cs_ext_list_price),
+           checksum(cs_ext_tax),
+           checksum(cs_coupon_amt),
+           checksum(cs_ext_ship_cost),
+           checksum(cs_net_paid),
+           checksum(cs_net_paid_inc_tax),
+           checksum(cs_net_paid_inc_ship),
+           checksum(cs_net_paid_inc_ship_tax),
+           checksum(cs_net_profit)
+    ] checksums
+from catalog_sales
+    )
+select checksum(cs) as table_checksum
+from column_checksums
+         cross join unnest(column_checksums.checksums) as x(cs);

benchmarks/tpc-ds/queries/refresh/verify_LF_I.sql

@@ -1,26 +1,18 @@
 -- ==============================================
---    Fetch two rows randomly from iv:
+-- 12015000
+-- a1 56 0d 3e 9a f7 5f 9d
 -- ==============================================
-with inv_random as (select
-                        inv_date_sk,
-                        inv_item_sk,
-                        inv_warehouse_sk,
-                        inv_quantity_on_hand
-                    from iv
-                    where inv_item_sk >= (select floor( max(inv_item_sk) * rand()) from iv )
-                    order by inv_item_sk limit 2)
+select count(*) from inventory;
 
--- ========================================================
---    Verify the row can be selected from inventory:
--- ========================================================
-
-select
-    inv.inv_date_sk,
-    inv.inv_item_sk,
-    inv.inv_warehouse_sk,
-    inv.inv_quantity_on_hand
-from inventory inv, inv_random
-where
-        inv.inv_date_sk = inv_random.inv_date_sk and
-        inv.inv_item_sk = inv_random.inv_item_sk and
-        inv.inv_warehouse_sk = inv_random.inv_warehouse_sk;
+with column_checksums as (
+    select array[
+               checksum(inv_date_sk),
+           checksum(inv_item_sk),
+           checksum(inv_warehouse_sk),
+           checksum(inv_quantity_on_hand)
+    ] checksums
+from inventory
+    )
+select checksum(cs) as table_checksum
+from column_checksums
+         cross join unnest(column_checksums.checksums) as x(cs);