问题
I'm using SQLite, and I have a table for properties, and a table for sub-properties. Each sub-property points to its parent using the fkPropertyId column. Right now, to create the initial database, I've got a script that looks something like this:
INSERT INTO property VALUES(1,.....);
INSERT INTO property VALUES(2,.....);
INSERT INTO property VALUES(3,.....);
INSERT INTO subproperty VALUES(1,.....,3);
INSERT INTO subproperty VALUES(2,.....,3);
INSERT INTO subproperty VALUES(3,.....,3);
INSERT INTO property VALUES(4,.....);
Now, I want to get rid of the hard-coded rowId, so it would be something like:
INSERT INTO property VALUES(NULL,.....);
INSERT INTO property VALUES(NULL,.....);
INSERT INTO property VALUES(NULL,.....);
INSERT INTO subproperty VALUES(NULL,.....,X);
INSERT INTO subproperty VALUES(NULL,.....,X);
INSERT INTO subproperty VALUES(NULL,.....,X);
INSERT INTO property VALUES(NULL,.....);
Where x refers to the last inserted rowId in the property table. Right now, that's
(SELECT MAX(rowId) FROM property)
Is there any better (and more technically accurate) way to write this script?
回答1:
Use the last_insert_rowid function:
SELECT last_insert_rowid();
回答2:
Well, the solution I came up with used the last_insert_rowid
function from Ben S:
INSERT INTO property VALUES(NULL,.....);
INSERT INTO property VALUES(NULL,.....);
INSERT INTO subproperty VALUES(1,.....,-1);
INSERT INTO subproperty VALUES(2,.....,-1);
INSERT INTO subproperty VALUES(3,.....,-1);
INSERT INTO property VALUES(NULL,.....);
UPDATE subproperty SET fkPropertyId = (SELECT last_insert_rowid()) WHERE fkPropertyId=-1;
INSERT INTO property VALUES(NULL,.....);
Not sure if that's the best approach, but it works for me, and it doesn't use any extra tables for temporary data storage.
来源:https://stackoverflow.com/questions/1822135/how-can-i-access-the-last-inserted-row-id-within-a-sql-script