Question

I use SQL at work to gather info, but nothing too advanced (yet).

I need to determine patterns of inventory instock and out of stock.

I have a table that shows item, location, inventory and date. I want to see if there is a pattern to the times when an item/location is out of stock and instock over time. The file contains a rolling three weeks of data. If an item is consistently going instock or out of stock in three weeks, then I need to know about it to research further.

An item/location is instock if inventory greater than zero. An item/location is out of stock if inventory is zero or negative.

Thanks for any assistance.

Sample data

Item, location, inventory, date
1243, 10,       2,         3/12/2012
1243, 10,       0,         3/13/2012
1243, 10,      -2,         3/14/2012
1243, 10,      -2,         3/15/2012
1243, 10,       4,         3/16/2012

Then additional item, location, inventory, date records.

Open to suggestions for output. I just need to see the item/location and how many times it switched between instock and out of stock in the three weeks worth of data.

Answer 1

Something like this may be what you’re looking for. This will count the number of times that each combination of item and location moves from “In Stock” to “Out of Stock” or vice versa. Note that the first row for each item and location, cannot be counted as a transition because we don’t know the prior status given this logic.

SELECT item,
       location,
       SUM( CASE WHEN status = 'In Stock' AND prior_status = 'Out of Stock'
                 THEN 1
                 ELSE 0
             END) moved_to_out_of_stock,
       SUM( CASE WHEN status = 'Out of Stock' AND prior_status = 'In Stock'
                 THEN 1
                 ELSE 0 
              END) moved_to_in_stock
  FROM (SELECT item, 
               location,
               status,
               lag( status ) over (partition by item, location
                                       order by dt) prior_status
          FROM (SELECT item, 
                       location, 
                       (case when inventory <= 0 
                             then 'Out of Stock'
                             else 'In Stock'
                         end) status,
                       dt
                  FROM your_table))
  GROUP BY item, location

This shows that the item went from Out of Stock to In Stock once and from In Stock to Out of Stock once in the sample data you posted.

SQL> ed
Wrote file afiedt.buf

  1  with your_table as (
  2    select 1243 item, 10 location, 2 inventory, date '2012-03-12' dt from dual union all
  3    select 1243, 10, 0, date '2012-03-13' from dual union all
  4    select 1243, 10, -2, date '2012-03-14' from dual union all
  5    select 1243, 10, -2, date '2012-03-15' from dual union all
  6    select 1243, 10, 4, date '2012-03-16' from dual
  7  )
  8  SELECT item,
  9         location,
 10         SUM( CASE WHEN status = 'In Stock' AND prior_status = 'Out of Stock'
 11                   THEN 1
 12                   ELSE 0
 13               END) moved_to_out_of_stock,
 14         SUM( CASE WHEN status = 'Out of Stock' AND prior_status = 'In Stock'
 15                   THEN 1
 16                   ELSE 0
 17                END) moved_to_in_stock
 18    FROM (SELECT item,
 19                 location,
 20                 status,
 21                 lag( status ) over (partition by item, location
 22                                         order by dt) prior_status
 23            FROM (SELECT item,
 24                         location,
 25                         (case when inventory <= 0
 26                               then 'Out of Stock'
 27                               else 'In Stock'
 28                           end) status,
 29                         dt
 30                    FROM your_table))
 31*   GROUP BY item, location
SQL> /

      ITEM   LOCATION MOVED_TO_OUT_OF_STOCK MOVED_TO_IN_STOCK
---------- ---------- --------------------- -----------------
      1243         10                     1                 1