Reputation: 4892
How I can in MySQL fetch any row by index from result set as it possible with arrays or collections in most programming languages ?
array[index]
Or:
collection.getElementByIndex(index)
I have a result set of dates, me need to check whether the 90 days between each date
Upvotes: 1
Views: 87
Reputation: 2075
You have two alternatives:
First alternative looks like:
SELECT BIT_AND(IFNULL(DATEDIFF((SELECT dt FROM foo WHERE dt > a.dt ORDER BY dt LIMIT 1), a.dt) >= 90, 1)) AS all_larger
FROM foo a;
Update: To handle a table where a date is duplicated, it is necessary to add a second sub-select to see if there are duplicates for the date, as follows:
SELECT BIT_AND(larger && ! duplicates) AS all_larger
FROM (SELECT a.dt
, IFNULL(DATEDIFF((SELECT dt FROM foo WHERE dt > a.dt ORDER BY dt LIMIT 1), a.dt) >= 90, 1) AS larger
, (SELECT COUNT(*) FROM foo WHERE dt = a.dt) > 1 AS duplicates
FROM foo a) AS x;
Second alternative looks like:
SET @prev = NULL;
SELECT BIT_AND(a.larger) AS all_larger
FROM (SELECT IFNULL(DATEDIFF(dt, @prev) >= 90, 1) AS larger
, @prev := dt
FROM foo ORDER BY dt) a;
Both give the following result set when run on a table where the difference between the dates are more than 90 days:
+------------+
| all_larger |
+------------+
| 1 |
+------------+
The second one is probably faster, but I haven't measured on larger sets.
Upvotes: 1
Reputation: 3732
Intrinsically you cannot. A relational database doesn't preserve record order (or at least you can't rely on it, even if it temporarily stores record order). In this way it acts more like a hashmap or List than an array.
However if you want, you can add a field in the table - let's call it RowNum - that stores a row number, and you can query on that.
select * from Table where RowNum = %index%;
Upvotes: 0