SQL: Add multiple rows from one table to another when no data for that date in new table - sql

Question

I'm trying to update medical data from one table to another after switching from one system to another. We have two tables, for simplicity I'll make this a simple example. There are many columns in these tables in reality (not just 5).
Table1:
name, date, var1, var2, var3
Table2:
name, date, var1a, var2a, var3a
I want to transfer data from Table 1 to Table 2 for any rows where there isn't previous data for that date, where var1 = var1a, etc (same columns with different names).
I was trying to do something with a loop, but realized that may not be necessary.
I had gotten this far but keep wasn't sure if this was ok:
UPDATE Table2 VALUES (date, var1a, var2a, var3a)
SELECT date, var1, var2, var3 FROM Table1
Is that correct syntax so far? Or do I need to map the variables to translate var1 into var1a, etc?
How do I add a check to make sure I don't overwrite any data already in Table1? I don't want to add data if there is already data for that date/name combination.
Thanks!

Answer 1

You can INSERT into TABLE2 all values from TABLE1 that do not already exist in Table2:
INSERT INTO Table2 (date, var1a, var2a, var3a)
SELECT date, var1, var2, var3
FROM Table1 t1
WHERE NOT EXISTS (SELECT 1 FROM Table2 t2 WHERE t2.date = t1.date)
Already existing values are specified by comparing the date column. You can add any other predicates in the SELECT subquery of the NOT EXISTS expression to suit your needs.

Answer 2

You could use an update with a join. And you dont need to update the date column since that's what you are using to find the matches in the 2 tables.
Either you generate a dynamic query based on the empty/null valued columns, or you could do something like the below, which puts the same value in the column if it exists in table2 or else puts the corresponding value from table1.
The below approach requires less logic and easier to implement but will produce IO equivalent to updating the entire table.
update tbl2
set val1a=isnull(val1a,val1)
, val2a=isnull(val2a,val2)
, val3a=isnull(val3a.val3)
from table1 tbl1
inner join table2 tbl2
on tbl1.name=tbl2.name
and tbl1.date=tbl2.date
Considerations:
The approach requires less logic and easier to implement but will produce IOs equivalent to updating the entire table2. If you have a smallish table i would go with this approach.
If its a big table then you should look into building specific query sets to reduce IO

Answer 3

This code is tested in Access but something very similar should work in SQL Server 2012:
UPDATE Table2 RIGHT JOIN Table1 ON Table2.date = Table1.date
SET Table2.name = Table1.name, Table2.date = Table1.date, Table2.var1 = Table1.var1a, Table2.var2 = Table1.var2a, Table2.var3 = Table1.var3a
WHERE (Table2.date Is Null);
Explanation: this uses a right join so that within the query all the data from Table1 is present and where there is a matched date for Table2 that data is present too. We then ignore all cases where there is any data for Table2 and update the query in all other cases - the update in fact inserts new data into Table2.