Link to home
Start Free TrialLog in
Avatar of amp834
amp834

asked on

Java JDBC ResultSet.insertRow, getting autoincrement field values

Using JDBC's  java.sql.ResultSet, I write a new record and want to get the automatically-generated key for the record that was just written.  Is there a way to do that?


// my current codebase uses ResultSet to write to a table
stmt = con.createStatement(java.sql.ResultSet.TYPE_FORWARD_ONLY, java.sql.ResultSet.CONCUR_UPDATABLE);
rs = stmt.executeQuery("select * from " + strTableName + " where 0=1");
rs.moveToInsertRow();
// update some fields
rs.insertRow();
// now I want to fetch the value of the autoincrement field
 
 
 
// If I switch the code to the sample below, there is a way to get the keys
 
query = "INSERT INTO collection ( name , memo ) VALUES ( " + Criteria.escape(args0.getName()) + " , " + Criteria.escape(args0.getMemo()) + " )";         
statement.execute(query,Statement.RETURN_GENERATED_KEYS);
ResultSet rs = statement.getGeneratedKeys();
if (rs.next()) {
  collectionId=rs.getInt(1);
}

Open in new window

Avatar of CEHJ
CEHJ
Flag of United Kingdom of Great Britain and Northern Ireland image

You're already doing it - with getGeneratedKeys..?
Oh i see what you mean - you want to use the first technique. Have you tried just getting the auto-inc value from the ResultSet directly with rs.getInt("autoincfieldname")?
Avatar of amp834
amp834

ASKER

I tried rs.getInt("autoincfieldname"), there is no "current" rowset after the insert operation.
Try calling rs.last() before attempting to getInt
Avatar of amp834

ASKER

I already tried that.  Supposedly it works if the database is MySql, but not necessarily for other databases.  And if other processes inserted records, then it may return their stuff.
Yes, the db/driver would have to support it.
Avatar of amp834

ASKER

The db/driver "supporting it" is just a special way MySql works.  it's not a standard to "goto last record" to get the just inserted record.

So that leaves the original question again!  

Either I have to switch to the prepared-statement "insert", or find some link between the resultset and the corresponding update statement, if there is such a thing.  Perhaps JDBC definition/standard doesn't support getting the inserted keys from resultset.insertRow.

Let's see if other people have any experience with this!
what database are you using?
there isn't a standard jdbc way to do it, so without knowing the database there is no way to determine what would work.
Avatar of amp834

ASKER

I'm using Derby right now, but want to be able to support other databases as well, maybe MySql, SqlServer, Postgres, etc.

Perhaps a better solution would be to take the ResultSet and write it out using a prepared statement (then discard the 'insertRow' ResultSet).  I could get the assigned key reliably.  (I read a benchmark for Derby that writing using prepared statements is slower than using ResultSet).

I assume the code for ResultSet.insertRow would be dependent on each JDBC driver.  (It seems it could be built on top of JDBC's prepared statements; if so, I could look at the code to see if I could make something similar).

I read something vaguely about a RowSet, maybe a temporary buffer for storing field values.  If there is such a thing, maybe I could use that instead of inventing another wheel.



yes writing with a resultset is typically going to be slower, whats the motivation for using insertRow() in the first place?
Avatar of amp834

ASKER

Actually, I read that writing with ResultSet is slower.

I wanted to use ResultSet because a lot of the work was already done, with the field names, field list, setting and getting, etc.,  but I suppose I can reinvent another wheel.  Many other people should have come across this issue, I wonder if there is a simple solution.

> Actually, I read that writing with ResultSet is slower.

thats what I said :)

Why aren't you using something like hibernate and save yourself all the overhead of implement the jdbc yourself.
Avatar of amp834

ASKER

I'm sorry, I meant writing with prepared statements was slower than with resultset (by about 20 to 30 %), but it seems you read the opposite!

I can look into using hibernate.  The reasons I stayed away:  
1. another layer to learn,
2. not sure if it can do things like scroll through a result set
3. not sure about its overhead/speed.  

For most simple gui operations, the speed doesn't matter.  But for batch processing of 500k+ records, it may.

4. I assume that hibernate would take care of getting the autoincrement value?

ASKER CERTIFIED SOLUTION
Avatar of Mick Barry
Mick Barry
Flag of Australia image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of amp834

ASKER

I will wait to see if there are any other comments and will close this in a week.  (My Hibernate books should be in by then!)

By the way, objects, I have a question posted about refreshing tree nodes, if you get a chance, can you look at it?
Avatar of amp834

ASKER

Thanks for the help.  I will use Prepared Statements, and maybe hibernate in the future.