I have a column of country dialcode numbers I want to filter the prefixes into the left most dial code
This is the source column:
prefix
------
542
54299
374
37477
37493
37494
37498
37447
37455
3749
37410
297
29756
29759
2976开发者_开发技巧6
29769
29796
29799
29773
29774
297600
297622
247
61
61861
61862
61863
This is a example of the result I want. Can sql do this easily and how or is there a better method. Bear in mind there will be about 30k rows
significant prefix
----------------------
542 542
542 54299
374 374
374 37477
374 37493
374 37494
374 37498
374 37447
374 37455
374 3749
374 37410
297 297
297 29756
297 29759
297 29766
297 29769
297 29796
297 29799
297 29773
297 29774
297 297600
297 297622
247 247
61 61
61 61861
61 61862
61 61863
You may want to try the following (using MySQL's INSTR()
and LENGTH()
functions):
SELECT ( SELECT prefix
FROM numbers n2
WHERE INSTR(n1.prefix, n2.prefix) = 1
ORDER BY LENGTH(n2.prefix)
LIMIT 1
) AS significant,
n1.prefix
FROM numbers n1;
Check out @onedaywhen's answer for an ANSI SQL version of the above query.
Test case:
CREATE TABLE numbers (prefix int);
INSERT INTO numbers VALUES (542);
INSERT INTO numbers VALUES (54299);
INSERT INTO numbers VALUES (374);
INSERT INTO numbers VALUES (37477);
INSERT INTO numbers VALUES (37493);
INSERT INTO numbers VALUES (37494);
INSERT INTO numbers VALUES (37498);
INSERT INTO numbers VALUES (37447);
INSERT INTO numbers VALUES (37455);
INSERT INTO numbers VALUES (3749);
INSERT INTO numbers VALUES (37410);
INSERT INTO numbers VALUES (297);
INSERT INTO numbers VALUES (29756);
INSERT INTO numbers VALUES (29759);
INSERT INTO numbers VALUES (29766);
INSERT INTO numbers VALUES (29769);
INSERT INTO numbers VALUES (29796);
INSERT INTO numbers VALUES (29799);
INSERT INTO numbers VALUES (29773);
INSERT INTO numbers VALUES (29774);
INSERT INTO numbers VALUES (297600);
INSERT INTO numbers VALUES (297622);
INSERT INTO numbers VALUES (247);
INSERT INTO numbers VALUES (61);
INSERT INTO numbers VALUES (61861);
INSERT INTO numbers VALUES (61862);
INSERT INTO numbers VALUES (61863);
Result:
+-------------+--------+
| significant | prefix |
+-------------+--------+
| 542 | 542 |
| 542 | 54299 |
| 374 | 374 |
| 374 | 37477 |
| 374 | 37493 |
| 374 | 37494 |
| 374 | 37498 |
| 374 | 37447 |
| 374 | 37455 |
| 374 | 3749 |
| 374 | 37410 |
| 297 | 297 |
| 297 | 29756 |
| 297 | 29759 |
| 297 | 29766 |
| 297 | 29769 |
| 297 | 29796 |
| 297 | 29799 |
| 297 | 29773 |
| 297 | 29774 |
| 297 | 297600 |
| 297 | 297622 |
| 247 | 247 |
| 61 | 61 |
| 61 | 61861 |
| 61 | 61862 |
| 61 | 61863 |
+-------------+--------+
27 rows in set (0.00 sec)
It should work even if you are using a varchar
to store the numbers.
UPDATE:
As for performance, you may want to consider caching the significant
part in your table:
CREATE TABLE numbers (prefix int, significant int);
-- Fill in the prefixes, leaving the significant field as NULL.
Then you could generate the significant
field as follows (using MySQL):
UPDATE numbers n
JOIN ( SELECT ( SELECT prefix
FROM numbers n2
WHERE INSTR(n1.prefix, n2.prefix) = 1
ORDER BY LENGTH(n2.prefix)
LIMIT 1
) AS significant,
n1.prefix
FROM numbers n1
) s ON (s.prefix = n.prefix)
SET n.significant = s.significant;
SELECT * FROM numbers;
+--------+-------------+
| prefix | significant |
+--------+-------------+
| 542 | 542 |
| 54299 | 542 |
| 374 | 374 |
| 37477 | 374 |
| 37493 | 374 |
| 37494 | 374 |
...
You would probably want to run the UPDATE
query whenever you add a new row in your numbers
table.
This question has only a 'sql' tag (no 'MySQL' tag) so here's a proposed solution using Standard SQL, assuming prefix
is an INTEGER
column in a table named Numbers
:
SELECT (
SELECT MIN(N2.prefix)
FROM Numbers AS N2
WHERE CAST(N1.prefix AS VARCHAR) LIKE CAST(N2.prefix AS VARCHAR) + '%'
) AS significant,
N1.prefix
FROM Numbers AS N1;
精彩评论