使用 INSERT INTO 和子查询 SELECT 插入字段的前一个值

Question

I have the following schema:

CREATE TABLE IF NOT EXISTS test (id INTEGER PRIMARY KEY, val INTEGER, previousval INTEGER)

I want to store the value val and automatically copy the previous value into the column previousval :

id  val previousval
1   100 NULL
2   200 100
3   300 200

I tried the following query and SQLite gives me an error message I don't understand.

INSERT INTO test as test1 (id, val, previousval) VALUES (NULL, 1, (SELECT test2.val FROM test AS test2 WHERE test2.id<test1.id ORDER BY test2.id DESC LIMIT 1))
> no such column: test1.id

I tried another query:

INSERT INTO test (id, val, previousval) VALUES (NULL, 100, (SELECT test2.val FROM test AS test2 WHERE test2.id<id ORDER BY test2.id DESC LIMIT 1)) 
INSERT INTO test (id, val, previousval) VALUES (NULL, 200, (SELECT test2.val FROM test AS test2 WHERE test2.id<id ORDER BY test2.id DESC LIMIT 1)) 
INSERT INTO test (id, val, previousval) VALUES (NULL, 300, (SELECT test2.val FROM test AS test2 WHERE test2.id<id ORDER BY test2.id DESC LIMIT 1)) 

The result is not what I expect and I don't understand why.

id  val previousval
1   100 NULL
2   200 NULL
3   300 NULL

I use sqlite-amalgamation-3310100. I also use SQLiteStudio for testing.

How to automatically copy the value of a field into a column using INSERT INTO ?

Answer 1

The problem with your attempt is that you are trying to correlated the subquery with a value that is not yet inserted. That can't work.

You could just do:

insert into test as test1 (id, val, previousval) 
values(
    null, 
    1, 
    (select val from test order by id desc limit 1)
)

Note: if id is an auto-incremented key (as it looks like), then just skip that column when inserting:

insert into test as test1 (val, previousval) 
values(
    1, 
    (select val from test order by id desc limit 1)
)

Bottom line, I would not actually recommend storing that derived information in the table itself: for example, if val is updated, then you need some additional processing to reset the following record.

You can easily compute that information on the fly when needed, or put in a view. If your version of SQLite supports window functions:

select id, val, lag(val) over(order by id) previousval from mytable

In earlier versions:

select
    t.id,
    t.val,
    (select val from mytable t1 where t1.id < t.id order by t1.id desc limit 1) previousval
from mytable t

Answer 2

In your CREATE statement of the table you define the column id as:

id INTEGER PRIMARY KEY 

and this means that it will be unique and auto incremented.
But this does not mean that it will always increase.
If you delete a row, the next time that you will insert a new row the missing id value may be reused, so the logic that you want to apply will not work.
So as it is explained here , you must disable the reuse of ids and this can be done by using the keyword AUTOINCREMENT in the definition of the column:

CREATE TABLE IF NOT EXISTS test (id INTEGER PRIMARY KEY AUTOINCREMENT, ...) 

Then you can insert new rows like this:

INSERT INTO test(val, previousval) VALUES
(100, (SELECT val FROM test ORDER BY id DESC LIMIT 1));
INSERT INTO test(val, previousval) VALUES       
(200, (SELECT val FROM test ORDER BY id DESC LIMIT 1));