In Chapter 4, you saw how the GROUP BY clause, along with the aggregate functions, can be used to produce summary results. For example, if you want to print the monthly total sales for each region, you would probably execute the following query:
SELECT r.name region, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY r.name, o.month; REGION MONTH SUM(O.TOT_SALES) -------------------- --------- ---------------- New England January 1527645 New England February 1847238 New England March 1699449 New England April 1792866 New England May 1698855 New England June 1510062 New England July 1678002 New England August 1642968 New England September 1726767 New England October 1648944 New England November 1384185 New England December 1599942 Mid-Atlantic January 1832091 Mid-Atlantic February 1286028 Mid-Atlantic March 1911093 Mid-Atlantic April 1623438 Mid-Atlantic May 1778805 Mid-Atlantic June 1504455 Mid-Atlantic July 1820742 Mid-Atlantic August 1381560 Mid-Atlantic September 1178694 Mid-Atlantic October 1530351 Mid-Atlantic November 1598667 Mid-Atlantic December 1477374 Southeast US January 1137063 Southeast US February 1855269 Southeast US March 1967979 Southeast US April 1830051 Southeast US May 1983282 Southeast US June 1705716 Southeast US July 1670976 Southeast US August 1436295 Southeast US September 1905633 Southeast US October 1610523 Southeast US November 1661598 Southeast US December 1841100 36 rows selected.
As expected, this report prints the total sales for each region and month combination. However, in a more complex application, you may also want to have the subtotal for each region over all months, along with the total for all regions, or you may want the subtotal for each month over all regions, along with the total for all months. In short, you may need to generate subtotals and totals at more than one level.
In data warehouse applications, you frequently need to generate summary information over various dimensions, and subtotal and total across those dimensions. Generating and retrieving this type of summary information is a core goal of almost all data warehouse applications.
By this time, you have realized that a simple GROUP BY query is not sufficient to generate the subtotals and totals described in this section. To illustrate the complexity of the problem, let's attempt to write a query that would return the following in a single output:
Sales for each month for every region
Subtotals over all months for every region
Total sales for all regions over all months
One way to generate multiple levels of summary (the only way prior to Oracle8i) is to write a UNION query. For example, the following UNION query will give us the desired three levels of subtotals:
SELECT r.name region, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY r.name, o.month UNION ALL SELECT r.name region, NULL, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY r.name UNION ALL SELECT NULL, NULL, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id; REGION MONTH SUM(O.TOT_SALES) -------------------- --------- ---------------- New England January 1527645 New England February 1847238 New England March 1699449 New England April 1792866 New England May 1698855 New England June 1510062 New England July 1678002 New England August 1642968 New England September 1726767 New England October 1648944 New England November 1384185 New England December 1599942 Mid-Atlantic January 1832091 Mid-Atlantic February 1286028 Mid-Atlantic March 1911093 Mid-Atlantic April 1623438 Mid-Atlantic May 1778805 Mid-Atlantic June 1504455 Mid-Atlantic July 1820742 Mid-Atlantic August 1381560 Mid-Atlantic September 1178694 Mid-Atlantic October 1530351 Mid-Atlantic November 1598667 Mid-Atlantic December 1477374 Southeast US January 1137063 Southeast US February 1855269 Southeast US March 1967979 Southeast US April 1830051 Southeast US May 1983282 Southeast US June 1705716 Southeast US July 1670976 Southeast US August 1436295 Southeast US September 1905633 Southeast US October 1610523 Southeast US November 1661598 Southeast US December 1841100 Mid-Atlantic 18923298 New England 19756923 Southeast US 20605485 59285706 40 rows selected.
This query produced 40 rows of output, 36 of which are the sales for each month for every region. The last four rows are the subtotals and the total. The three rows with region names and NULL values for the month are the subtotals for each region over all the months, and the last row with NULL values for both the region and month is the total sales for all the regions over all the months.
Now that you have the desired result, try to analyze the query a bit. You have a very small all_orders table with only 1440 rows in this example. You wanted to have summary information over just two dimensions?region and month. You have 3 regions and 12 months. To get the desired summary information from this table, you have to write a query consisting of three SELECT statements combined together using UNION ALL. The execution plan for this query is:
PLAN_TABLE_OUTPUT ----------------------------------------------------- ----------------------------------------------------- | Id | Operation | Name | ----------------------------------------------------- | 0 | SELECT STATEMENT | | | 1 | UNION-ALL | | | 2 | SORT GROUP BY | | | 3 | MERGE JOIN | | | 4 | TABLE ACCESS BY INDEX ROWID| REGION | | 5 | INDEX FULL SCAN | REGION_PK | |* 6 | SORT JOIN | | | 7 | TABLE ACCESS FULL | ALL_ORDERS | | 8 | SORT GROUP BY | | | 9 | MERGE JOIN | | | 10| TABLE ACCESS BY INDEX ROWID| REGION | | 11| INDEX FULL SCAN | REGION_PK | |* 12| SORT JOIN | | | 13| TABLE ACCESS FULL | ALL_ORDERS | | 14| SORT AGGREGATE | | | 15| NESTED LOOPS | | | 16| TABLE ACCESS FULL | ALL_ORDERS | |* 17| INDEX UNIQUE SCAN | REGION_PK | -----------------------------------------------------
As indicated by the execution plan output, Oracle needs to perform the following operations to get the results:
Three FULL TABLE scans on all_orders Three INDEX scan on region_pk (Primary key of table region) Two Sort-Merge Joins One NESTED LOOPS JOIN Two SORT GROUP BY operations One SORT AGGREGATE operation One UNION ALL
In any practical application the all_orders table will consist of millions of rows, and performing all these operations would be time-consuming. Even worse, if you have more dimensions for which to prepare summary information than the two shown in this example, you have to write an even more complex query. The bottom line is that such a query badly hurts performance.
Oracle8i introduced several new features for generating multiple levels of summary information with one query. One such feature is a set of extensions to the GROUP BY clause. In Oracle8i, the GROUP BY clause comes with two extensions: ROLLUP and CUBE. Oracle9i introduced another extension: GROUPING SETS. We discuss ROLLUP in this section. CUBE and GROUPING SETS are discussed later in this chapter.
ROLLUP is an extension to the GROUP BY clause, and therefore can only appear in a query with a GROUP BY clause. The ROLLUP operation groups the selected rows based on the expressions in the GROUP BY clause, and prepares a summary row for each group. The syntax of ROLLUP is:
SELECT . . . FROM . . . GROUP BY ROLLUP (ordered list of grouping columns)
Using ROLLUP, you can generate the summary information discussed in the previous section in a much easier way than in our UNION ALL query. For example:
SELECT r.name region, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY ROLLUP (r.name, o.month); REGION MONTH SUM(O.TOT_SALES) -------------------- --------- ---------------- New England January 1527645 New England February 1847238 New England March 1699449 New England April 1792866 New England May 1698855 New England June 1510062 New England July 1678002 New England August 1642968 New England September 1726767 New England October 1648944 New England November 1384185 New England December 1599942 New England 19756923 Mid-Atlantic January 1832091 Mid-Atlantic February 1286028 Mid-Atlantic March 1911093 Mid-Atlantic April 1623438 Mid-Atlantic May 1778805 Mid-Atlantic June 1504455 Mid-Atlantic July 1820742 Mid-Atlantic August 1381560 Mid-Atlantic September 1178694 Mid-Atlantic October 1530351 Mid-Atlantic November 1598667 Mid-Atlantic December 1477374 Mid-Atlantic 18923298 Southeast US January 1137063 Southeast US February 1855269 Southeast US March 1967979 Southeast US April 1830051 Southeast US May 1983282 Southeast US June 1705716 Southeast US July 1670976 Southeast US August 1436295 Southeast US September 1905633 Southeast US October 1610523 Southeast US November 1661598 Southeast US December 1841100 Southeast US 20605485 59285706 40 rows selected.
As you can see in this output, the ROLLUP operation produced subtotals across the specified dimensions and a grand total. The argument to the ROLLUP operation is an ordered list of grouping columns. Since the ROLLUP operation is used in conjunction with the GROUP BY clause, it first generates aggregate values based on the GROUP BY operation on the ordered list of columns. It then generates higher-level subtotals and finally a grand total. ROLLUP not only simplifies the query, it results in more efficient execution. The execution plan for this ROLLUP query is as follows:
PLAN_TABLE_OUTPUT ---------------------------------------------------- ---------------------------------------------------- | Id | Operation | Name | ---------------------------------------------------- | 0 | SELECT STATEMENT | | | 1 | SORT GROUP BY ROLLUP | | | 2 | MERGE JOIN | | | 3 | TABLE ACCESS BY INDEX ROWID| REGION | | 4 | INDEX FULL SCAN | REGION_PK | |* 5 | SORT JOIN | | | 6 | TABLE ACCESS FULL | ALL_ORDERS | ----------------------------------------------------
Rather than the multiple table scans, joins, and other operations required by the UNION ALL version of the query, the ROLLUP query needs just one index scan on region_pk, one full table scan on all_orders, and one join to generate the required output.
If you want to generate subtotals for each month instead of for each region, all you need to do is change the order of columns in the ROLLUP operation, as in the following example:
SELECT r.name region, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY ROLLUP (o.month, r.name); REGION MONTH SUM(O.TOT_SALES) -------------------- --------- ---------------- New England January 1527645 Mid-Atlantic January 1832091 Southeast US January 1137063 January 4496799 New England February 1847238 Mid-Atlantic February 1286028 Southeast US February 1855269 February 4988535 New England March 1699449 Mid-Atlantic March 1911093 Southeast US March 1967979 March 5578521 New England April 1792866 Mid-Atlantic April 1623438 Southeast US April 1830051 April 5246355 New England May 1698855 Mid-Atlantic May 1778805 Southeast US May 1983282 May 5460942 New England June 1510062 Mid-Atlantic June 1504455 Southeast US June 1705716 June 4720233 New England July 1678002 Mid-Atlantic July 1820742 Southeast US July 1670976 July 5169720 New England August 1642968 Mid-Atlantic August 1381560 Southeast US August 1436295 August 4460823 New England September 1726767 Mid-Atlantic September 1178694 Southeast US September 1905633 September 4811094 New England October 1648944 Mid-Atlantic October 1530351 Southeast US October 1610523 October 4789818 New England November 1384185 Mid-Atlantic November 1598667 Southeast US November 1661598 November 4644450 New England December 1599942 Mid-Atlantic December 1477374 Southeast US December 1841100 December 4918416 59285706 49 rows selected.
Adding dimensions does not result in additional complexity. The following query rolls up subtotals for the region, the month, and the year for the first quarter:
SELECT o.year, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, r.name region, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id WHERE o.month BETWEEN 1 AND 3 GROUP BY ROLLUP (o.year, o.month, r.name); YEAR MONTH REGION SUM(O.TOT_SALES) ---------- --------- -------------------- ---------------- 2000 January New England 1018430 2000 January Mid-Atlantic 1221394 2000 January Southeast US 758042 2000 January 2997866 2000 February New England 1231492 2000 February Mid-Atlantic 857352 2000 February Southeast US 1236846 2000 February 3325690 2000 March New England 1132966 2000 March Mid-Atlantic 1274062 2000 March Southeast US 1311986 2000 March 3719014 2000 10042570 2001 January New England 509215 2001 January Mid-Atlantic 610697 2001 January Southeast US 379021 2001 January 1498933 2001 February New England 615746 2001 February Mid-Atlantic 428676 2001 February Southeast US 618423 2001 February 1662845 2001 March New England 566483 2001 March Mid-Atlantic 637031 2001 March Southeast US 655993 2001 March 1859507 2001 5021285 15063855 27 rows selected.
In a ROLLUP query with N dimensions, the grand total is considered the top level. The various subtotal rows of N-1 dimensions constitute the next lower level, the subtotal rows of N-2 dimensions constitute yet another level down, and so on. In the most recent example, you have three dimensions (year, month, and region), and the total row is the top level. The subtotal rows for the year represent the next lower level, because these rows are subtotals across two dimensions (month and region). The subtotal rows for the year and month combination are one level lower, because these rows are subtotals across one dimension (region). The rest of the rows are the result of the regular GROUP BY operation (without ROLLUP), and form the lowest level.
If you want to exclude some subtotals and totals from the ROLLUP output, you can only move top to bottom, i.e., exclude the top-level total first, then progressively go down to the next level subtotals, and so on. To do this, you have to take out one or more columns from the ROLLUP operation, and put them in the GROUP BY clause. This is called a partial ROLLUP.
As an example of a partial ROLLUP, let's see what happens when you take out the first column, which is o.year, from the previous ROLLUP operation and move it into the GROUP BY clause.
SELECT o.year, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, r.name region, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id WHERE o.month BETWEEN 1 AND 3 GROUP BY o.year, ROLLUP (o.month, r.name); YEAR MONTH REGION SUM(O.TOT_SALES) ---------- --------- -------------------- ---------------- 2000 January New England 1018430 2000 January Mid-Atlantic 1221394 2000 January Southeast US 758042 2000 January 2997866 2000 February New England 1231492 2000 February Mid-Atlantic 857352 2000 February Southeast US 1236846 2000 February 3325690 2000 March New England 1132966 2000 March Mid-Atlantic 1274062 2000 March Southeast US 1311986 2000 March 3719014 2000 10042570 2001 January New England 509215 2001 January Mid-Atlantic 610697 2001 January Southeast US 379021 2001 January 1498933 2001 February New England 615746 2001 February Mid-Atlantic 428676 2001 February Southeast US 618423 2001 February 1662845 2001 March New England 566483 2001 March Mid-Atlantic 637031 2001 March Southeast US 655993 2001 March 1859507 2001 5021285 26 rows selected.
The query in this example excludes the grand-total row from the output. By taking out o.year from the ROLLUP operation, you are asking the database not to roll up summary information over the years. Therefore, the database rolls up summary information on region and month. When you proceed to remove o.month from the ROLLUP operation, the query will not generate the roll up summary for the month dimension, and only the region-level subtotals will be printed in the output. For example:
SELECT o.year, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, r.name region, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id WHERE o.month BETWEEN 1 AND 3 GROUP BY o.year, o.month, ROLLUP (r.name); YEAR MONTH REGION SUM(O.TOT_SALES) ---------- --------- -------------------- ---------------- 2000 January New England 1018430 2000 January Mid-Atlantic 1221394 2000 January Southeast US 758042 2000 January 2997866 2000 February New England 1231492 2000 February Mid-Atlantic 857352 2000 February Southeast US 1236846 2000 February 3325690 2000 March New England 1132966 2000 March Mid-Atlantic 1274062 2000 March Southeast US 1311986 2000 March 3719014 2001 January New England 509215 2001 January Mid-Atlantic 610697 2001 January Southeast US 379021 2001 January 1498933 2001 February New England 615746 2001 February Mid-Atlantic 428676 2001 February Southeast US 618423 2001 February 1662845 2001 March New England 566483 2001 March Mid-Atlantic 637031 2001 March Southeast US 655993 2001 March 1859507 24 rows selected.
The CUBE extension of the GROUP BY clause takes aggregation one step further than ROLLUP. The CUBE operation generates subtotals for all possible combinations of the grouping columns. Therefore, output of a CUBE operation will contain all subtotals produced by an equivalent ROLLUP operation and also some additional subtotals. For example, if you are performing ROLLUP on columns region and month, you will get subtotals for all months for each region, and a grand total. However, if you perform the corresponding CUBE, you will get:
The regular rows produced by the GROUP BY clause
Subtotals for all months on each region
A subtotal for all regions on each month
A grand total
Like ROLLUP, CUBE is an extension of the GROUP BY clause, and can appear in a query only along with a GROUP BY clause. The syntax of CUBE is:
SELECT . . . FROM . . . GROUP BY CUBE (list of grouping columns)
For example, the following query returns subtotals for all combinations of regions and months in the all_orders table:
SELECT r.name region, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY CUBE(r.name, o.month); REGION MONTH SUM(O.TOT_SALES) -------------------- --------- ---------------- 59285706 January 4496799 February 4988535 March 5578521 April 5246355 May 5460942 June 4720233 July 5169720 August 4460823 September 4811094 October 4789818 November 4644450 December 4918416 New England 19756923 New England January 1527645 New England February 1847238 New England March 1699449 New England April 1792866 New England May 1698855 New England June 1510062 New England July 1678002 New England August 1642968 New England September 1726767 New England October 1648944 New England November 1384185 New England December 1599942 Mid-Atlantic 18923298 Mid-Atlantic January 1832091 Mid-Atlantic February 1286028 Mid-Atlantic March 1911093 Mid-Atlantic April 1623438 Mid-Atlantic May 1778805 Mid-Atlantic June 1504455 Mid-Atlantic July 1820742 Mid-Atlantic August 1381560 Mid-Atlantic September 1178694 Mid-Atlantic October 1530351 Mid-Atlantic November 1598667 Mid-Atlantic December 1477374 Southeast US 20605485 Southeast US January 1137063 Southeast US February 1855269 Southeast US March 1967979 Southeast US April 1830051 Southeast US May 1983282 Southeast US June 1705716 Southeast US July 1670976 Southeast US August 1436295 Southeast US September 1905633 Southeast US October 1610523 Southeast US November 1661598 Southeast US December 1841100 52 rows selected.
Note that the output contains not only the subtotals for each region, but also the subtotals for each month. You can get the same result from a query without the CUBE operation. However, that query would be lengthy and complex and, of course, very inefficient. Such a query would look like:
SELECT NULL region, NULL month, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id UNION ALL SELECT NULL, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY o.month UNION ALL SELECT r.name region, NULL, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY r.name UNION ALL SELECT r.name region, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY r.name, o.month; REGION MONTH SUM(O.TOT_SALES) -------------------- --------- ---------------- 59285706 January 4496799 February 4988535 March 5578521 April 5246355 May 5460942 June 4720233 July 5169720 August 4460823 September 4811094 October 4789818 November 4644450 December 4918416 Mid-Atlantic 18923298 New England 19756923 Southeast US 20605485 New England January 1527645 New England February 1847238 New England March 1699449 New England April 1792866 New England May 1698855 New England June 1510062 New England July 1678002 New England August 1642968 New England September 1726767 New England October 1648944 New England November 1384185 New England December 1599942 Mid-Atlantic January 1832091 Mid-Atlantic February 1286028 Mid-Atlantic March 1911093 Mid-Atlantic April 1623438 Mid-Atlantic May 1778805 Mid-Atlantic June 1504455 Mid-Atlantic July 1820742 Mid-Atlantic August 1381560 Mid-Atlantic September 1178694 Mid-Atlantic October 1530351 Mid-Atlantic November 1598667 Mid-Atlantic December 1477374 Southeast US January 1137063 Southeast US February 1855269 Southeast US March 1967979 Southeast US April 1830051 Southeast US May 1983282 Southeast US June 1705716 Southeast US July 1670976 Southeast US August 1436295 Southeast US September 1905633 Southeast US October 1610523 Southeast US November 1661598 Southeast US December 1841100 52 rows selected.
Since a CUBE produces aggregate results for all possible combinations of the grouping columns, the output of a query using CUBE is independent of the order of columns in the CUBE operation, if everything else remains the same. This is not the case with ROLLUP. If everything else in the query remains the same, ROLLUP(a,b) will produce a slightly different result set than ROLLUP(b,a). However, the result set of CUBE(a,b) will be the same as that of CUBE(b,a). The following example illustrates this by taking the example at the beginning of this section and reversing the order of columns in the CUBE operation:
SELECT r.name region, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY CUBE(o.month, r.name); REGION MONTH SUM(O.TOT_SALES) -------------------- --------- ---------------- 59285706 New England 19756923 Mid-Atlantic 18923298 Southeast US 20605485 January 4496799 New England January 1527645 Mid-Atlantic January 1832091 Southeast US January 1137063 February 4988535 New England February 1847238 Mid-Atlantic February 1286028 Southeast US February 1855269 March 5578521 New England March 1699449 Mid-Atlantic March 1911093 Southeast US March 1967979 April 5246355 New England April 1792866 Mid-Atlantic April 1623438 Southeast US April 1830051 May 5460942 New England May 1698855 Mid-Atlantic May 1778805 Southeast US May 1983282 June 4720233 New England June 1510062 Mid-Atlantic June 1504455 Southeast US June 1705716 July 5169720 New England July 1678002 Mid-Atlantic July 1820742 Southeast US July 1670976 August 4460823 New England August 1642968 Mid-Atlantic August 1381560 Southeast US August 1436295 September 4811094 New England September 1726767 Mid-Atlantic September 1178694 Southeast US September 1905633 October 4789818 New England October 1648944 Mid-Atlantic October 1530351 Southeast US October 1610523 November 4644450 New England November 1384185 Mid-Atlantic November 1598667 Southeast US November 1661598 December 4918416 New England December 1599942 Mid-Atlantic December 1477374 Southeast US December 1841100 52 rows selected.
This query produced the same results as the earlier query; only the order of the rows happens to be different.
To exclude some subtotals from the output, you can do a partial CUBE, (similar to a partial ROLLUP) by taking out column(s) from the CUBE operation and putting them into the GROUP BY clause. Here's an example:
SELECT r.name region, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY r.name, CUBE(o.month); REGION MONTH SUM(O.TOT_SALES) -------------------- --------- ---------------- New England 19756923 New England January 1527645 New England February 1847238 New England March 1699449 New England April 1792866 New England May 1698855 New England June 1510062 New England July 1678002 New England August 1642968 New England September 1726767 New England October 1648944 New England November 1384185 New England December 1599942 Mid-Atlantic 18923298 Mid-Atlantic January 1832091 Mid-Atlantic February 1286028 Mid-Atlantic March 1911093 Mid-Atlantic April 1623438 Mid-Atlantic May 1778805 Mid-Atlantic June 1504455 Mid-Atlantic July 1820742 Mid-Atlantic August 1381560 Mid-Atlantic September 1178694 Mid-Atlantic October 1530351 Mid-Atlantic November 1598667 Mid-Atlantic December 1477374 Southeast US 20605485 Southeast US January 1137063 Southeast US February 1855269 Southeast US March 1967979 Southeast US April 1830051 Southeast US May 1983282 Southeast US June 1705716 Southeast US July 1670976 Southeast US August 1436295 Southeast US September 1905633 Southeast US October 1610523 Southeast US November 1661598 Southeast US December 1841100 39 rows selected.
If you compare the results of the partial CUBE operation with that of the full CUBE operation, discussed at the beginning of this section, you will notice that the partial CUBE has excluded the subtotals for each month and the grand total from the output. If you want to retain the subtotals for each month, but want to exclude the subtotals for each region, you can swap the position of r.name and o.month in the GROUP BY . . . CUBE clause, as shown here:
SELECT r.name region, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY o.month, CUBE(r.name);
One interesting thing to note is that if you have one column in the CUBE operation, it produces the same result as the ROLLUP operation. Therefore, the following two queries produce identical results:
SELECT r.name region, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY r.name, CUBE(o.month); SELECT r.name region, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id GROUP BY r.name, ROLLUP(o.month);
ROLLUP and CUBE produce extra rows in the output that contain subtotals and totals. When a row represents a summary over a given column or set of columns, those columns will contain NULL values. Output containing NULLs and indicating subtotals doesn't make sense to an ordinary person who is unware of the behavior of ROLLUP and CUBE operations. Does your corporate vice president (VP) care about whether you used ROLLUP or CUBE or any other operation to get him the monthly total sales for each region? Obviously, he doesn't. That's exactly why you are reading this page and not your VP.
If you know your way around the NVL function, you would probably attempt to translate each NULL value from CUBE and ROLLUP to some descriptive value, as in the following example:
SELECT NVL(TO_CHAR(o.year), 'All Years') year, NVL(TO_CHAR(TO_DATE(o.month, 'MM'), 'Month'), 'First Quarter') month, NVL(r.name, 'All Regions') region, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id WHERE o.month BETWEEN 1 AND 3 GROUP BY ROLLUP (o.year, o.month, r.name); YEAR MONTH REGION SUM(O.TOT_SALES) ------------ ------------- -------------- ---------------- 2000 January New England 1018430 2000 January Mid-Atlantic 1221394 2000 January Southeast US 758042 2000 January All Regions 2997866 2000 February New England 1231492 2000 February Mid-Atlantic 857352 2000 February Southeast US 1236846 2000 February All Regions 3325690 2000 March New England 1132966 2000 March Mid-Atlantic 1274062 2000 March Southeast US 1311986 2000 March All Regions 3719014 2000 First Quarter All Regions 10042570 2001 January New England 509215 2001 January Mid-Atlantic 610697 2001 January Southeast US 379021 2001 January All Regions 1498933 2001 February New England 615746 2001 February Mid-Atlantic 428676 2001 February Southeast US 618423 2001 February All Regions 1662845 2001 March New England 566483 2001 March Mid-Atlantic 637031 2001 March Southeast US 655993 2001 March All Regions 1859507 2001 First Quarter All Regions 5021285 All Years First Quarter All Regions 15063855 27 rows selected.
The NVL function works pretty well for this example. However, if the data itself contains some NULL values, it becomes impossible to distinguish whether a NULL value represents unavailable data or a subtotal row. The NVL function will cause a problem in such a case. The following data can be used to illustrate this problem:
SELECT * FROM disputed_orders; ORDER_NBR CUST_NBR SALES_EMP_ID SALE_PRICE ORDER_DT EXPECTED_ STATUS ---------- ---------- ------------ ---------- --------- --------- --------- 1001 1 7354 99 22-JUL-01 23-JUL-01 DELIVERED 1000 1 7354 19-JUL-01 24-JUL-01 1002 5 7368 12-JUL-01 25-JUL-01 1003 4 7654 56 16-JUL-01 26-JUL-01 DELIVERED 1004 4 7654 34 18-JUL-01 27-JUL-01 PENDING 1005 8 7654 99 22-JUL-01 24-JUL-01 DELIVERED 1006 1 7354 22-JUL-01 28-JUL-01 1007 5 7368 25 20-JUL-01 22-JUL-01 PENDING 1008 5 7368 25 21-JUL-01 23-JUL-01 PENDING 1009 1 7354 56 18-JUL-01 22-JUL-01 DELIVERED 1012 1 7354 99 22-JUL-01 23-JUL-01 DELIVERED 1011 1 7354 19-JUL-01 24-JUL-01 1015 5 7368 12-JUL-01 25-JUL-01 1017 4 7654 56 16-JUL-01 26-JUL-01 DELIVERED 1019 4 7654 34 18-JUL-01 27-JUL-01 PENDING 1021 8 7654 99 22-JUL-01 24-JUL-01 DELIVERED 1023 1 7354 22-JUL-01 28-JUL-01 1025 5 7368 25 20-JUL-01 22-JUL-01 PENDING 1027 5 7368 25 21-JUL-01 23-JUL-01 PENDING 1029 1 7354 56 18-JUL-01 22-JUL-01 DELIVERED 20 rows selected.
Note that the column status contains NULL values. If you want the summary status of orders for each customer, and you executed the following query (note the application of NVL to the status column), the output might surprise you.
SELECT NVL(TO_CHAR(cust_nbr), 'All Customers') customer, NVL(status, 'All Status') status, COUNT(*) FROM disputed_orders GROUP BY CUBE(cust_nbr, status); CUSTOMER STATUS COUNT(*) ---------------------------------------- -------------------- ---------- All Customers All Status 6 All Customers All Status 20 All Customers PENDING 6 All Customers DELIVERED 8 1 All Status 4 1 All Status 8 1 DELIVERED 4 4 All Status 4 4 PENDING 2 4 DELIVERED 2 5 All Status 2 5 All Status 6 5 PENDING 4 8 All Status 2 8 DELIVERED 2 15 rows selected.
This output doesn't make any sense. The problem is that any time the status column legitimately contains a NULL value, the NVL function returns the string "All Status." Obviously, NVL isn't useful in this situation. However, don't worry?Oracle provides a solution to this problem through the GROUPING function.
The GROUPING function is meant to be used in conjunction with either a ROLLUP or a CUBE operation. The GROUPING function takes a grouping column name as input and returns either 1 or 0. A 1 is returned if the column's value is NULL as the result of aggregation (ROLLUP or CUBE); otherwise, 0 is returned. The general syntax of the GROUPING function is:
SELECT . . . [GROUPING(grouping_column_name)] . . . FROM . . . GROUP BY . . . {ROLLUP | CUBE} (grouping_column_name)
The following example illustrates the use of GROUPING function in a simple way by returning the GROUPING function results for the three columns passed to ROLLUP:
SELECT o.year, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, r.name region, SUM(o.tot_sales), GROUPING(o.year) y, GROUPING(o.month) m, GROUPING(r.name) r FROM all_orders o JOIN region r ON r.region_id = o.region_id WHERE o.month BETWEEN 1 AND 3 GROUP BY ROLLUP (o.year, o.month, r.name); YEAR MONTH REGION SUM(O.TOT_SALES) Y M R ----- --------- -------------------- ---------------- ---- ----- ----- 2000 January New England 1018430 0 0 0 2000 January Mid-Atlantic 1221394 0 0 0 2000 January Southeast US 758042 0 0 0 2000 January 2997866 0 0 1 2000 February New England 1231492 0 0 0 2000 February Mid-Atlantic 857352 0 0 0 2000 February Southeast US 1236846 0 0 0 2000 February 3325690 0 0 1 2000 March New England 1132966 0 0 0 2000 March Mid-Atlantic 1274062 0 0 0 2000 March Southeast US 1311986 0 0 0 2000 March 3719014 0 0 1 2000 10042570 0 1 1 2001 January New England 509215 0 0 0 2001 January Mid-Atlantic 610697 0 0 0 2001 January Southeast US 379021 0 0 0 2001 January 1498933 0 0 1 2001 February New England 615746 0 0 0 2001 February Mid-Atlantic 428676 0 0 0 2001 February Southeast US 618423 0 0 0 2001 February 1662845 0 0 1 2001 March New England 566483 0 0 0 2001 March Mid-Atlantic 637031 0 0 0 2001 March Southeast US 655993 0 0 0 2001 March 1859507 0 0 1 2001 5021285 0 1 1 15063855 1 1 1 27 rows selected.
Look at the y, m, and r columns in this output. Row 4 is a region-level subtotal for a particular month and year, and therefore, the GROUPING function results in a value of 1 for the region and a value 0 for the month and year. Row 26 (the second to last) is a subtotal for all regions and months for a particular year, and therefore, the GROUPING function prints 1 for the month and the region and 0 for the year. Row 27 (the grand total) contains 1 for all the GROUPING columns.
With a combination of GROUPING and DECODE (or CASE), you can produce more readable query output when using CUBE and ROLLUP, as in the following example:
SELECT DECODE(GROUPING(o.year), 1, 'All Years', o.year) Year, DECODE(GROUPING(o.month), 1, 'All Months', TO_CHAR(TO_DATE(o.month, 'MM'), 'Month')) Month, DECODE(GROUPING(r.name), 1, 'All Regions', r.name) Region, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id WHERE o.month BETWEEN 1 AND 3 GROUP BY ROLLUP (o.year, o.month, r.name); YEAR MONTH REGION SUM(O.TOT_SALES) ---------------- ---------- -------------------- ---------------- 2000 January New England 1018430 2000 January Mid-Atlantic 1221394 2000 January Southeast US 758042 2000 January All Regions 2997866 2000 February New England 1231492 2000 February Mid-Atlantic 857352 2000 February Southeast US 1236846 2000 February All Regions 3325690 2000 March New England 1132966 2000 March Mid-Atlantic 1274062 2000 March Southeast US 1311986 2000 March All Regions 3719014 2000 All Months All Regions 10042570 2001 January New England 509215 2001 January Mid-Atlantic 610697 2001 January Southeast US 379021 2001 January All Regions 1498933 2001 February New England 615746 2001 February Mid-Atlantic 428676 2001 February Southeast US 618423 2001 February All Regions 1662845 2001 March New England 566483 2001 March Mid-Atlantic 637031 2001 March Southeast US 655993 2001 March All Regions 1859507 2001 All Months All Regions 5021285 All Years All Months All Regions 15063855 27 rows selected.
By using DECODE with GROUPING, we produced the same result that was produced by using NVL at the beginning of the section. However, the risk of mistreating a NULL data value as a summary row is eliminated by using GROUPING and DECODE. You will notice this in the following example, in which NULL data values in subtotal and total rows are treated differently by the GROUPING function than the NULL values in the summary rows:
SELECT DECODE(GROUPING(cust_nbr), 1, 'All Customers', cust_nbr) customer, DECODE(GROUPING(status), 1, 'All Status', status) status, COUNT(*) FROM disputed_orders GROUP BY CUBE(cust_nbr, status); CUSTOMER STATUS COUNT(*) ---------------------------------------- -------------------- ---------- All Customers 6 All Customers All Status 20 All Customers PENDING 6 All Customers DELIVERED 8 1 4 1 All Status 8 1 DELIVERED 4 4 All Status 4 4 PENDING 2 4 DELIVERED 2 5 2 5 All Status 6 5 PENDING 4 8 All Status 2 8 DELIVERED 2 15 rows selected.
Earlier in this chapter, you saw how to generate summary information using ROLLUP and CUBE. However, the output of ROLLUP and CUBE include the rows produced by the regular GROUP BY operation along with the summary rows. Oracle9i introduced another extension to the GROUP BY clause called GROUPING SETS that you can use to generate summary information at the level you choose without including all the rows produced by the regular GROUP BY operation.
Like ROLLUP and CUBE, GROUPING SETS is also an extension of the GROUP BY clause, and can appear in a query only along with a GROUP BY clause. The syntax of GROUPING SETS is:
SELECT . . . FROM . . . GROUP BY GROUPING SETS (list of grouping columns)
Let's take an example to understand the GROUPING SETS operation further:
SELECT o.year, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, r.name region, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id WHERE o.month BETWEEN 1 AND 3 GROUP BY GROUPING SETS (o.year, o.month, r.name); YEAR MONTH REGION SUM(O.TOT_SALES) ---------- --------- -------------------- ---------------- Mid-Atlantic 5029212 New England 5074332 Southeast US 4960311 January 4496799 February 4988535 March 5578521 2000 10042570 2001 5021285 8 rows selected.
This output contains only the subtotals at the region, month, and year levels, but that none of the normal, more detailed, GROUP BY data is included. The order of columns in the GROUPING SETS operation is not critical. The operation produces the same output regardless of the order of the columns. For example:
SELECT o.year, TO_CHAR(TO_DATE(o.month, 'MM'), 'Month') month, r.name region, SUM(o.tot_sales) FROM all_orders o JOIN region r ON r.region_id = o.region_id WHERE o.month BETWEEN 1 AND 3 GROUP BY GROUPING SETS (o.month, r.name, o.year); YEAR MONTH REGION SUM(O.TOT_SALES) ---------- --------- -------------------- ---------------- Mid-Atlantic 5029212 New England 5074332 Southeast US 4960311 January 4496799 February 4988535 March 5578521 2000 10042570 2001 5021285 8 rows selected.