On Tue, Aug 25, 2009 at 10:56 AM, Naftoli Gugenheim <[email protected]>wrote:
> It definitely would have been an option, although it would have been
> more difficult due to the complex structure of the new database. But I
> want to understand why it can't work the way I thought it would.
Peter provided code that allows you to override autogeneration of keys.
Why does it work the way it does? Because in the nearly 3 year life of
Mapper, you're the first person that's asked for the feature. Because the
cost of changing Mapper to work differently (especially when there's a
work-around) and testing it and making sure it doesn't break isn't worth it.
>
> Anyway, here's the code that I used in the end. There may still be a
> few bugs but it demonstrates the approach. I attached the whole file
> but here's a little sample.
> val clients = scala.collection.mutable.Map[Long, Client]()
> ...
> DB.exec(sourceCon, "select * from clients") {rs =>
> import rs._
> while(rs.next) {
> val c =
> Client.create.last(getString("last")).first(getString("first")).
>
>
> home(getString("home")).cell(getString("cell")).address(getString("address")).
>
>
> area(getString("area")).created(getDate("created")).modified(getDate("modified"))
> if(!c.save)
> println("Couldn't save: " + c)
> clients += getLong("id") -> c
> }
> }
> Then I have a Map from the old id to the Client, and when I process
> the Request which has a client_id I can write
> req.client(clients(getLong("client_id")))
>
> So I guess I could do this by writing direct SQL, but then after I
> wrote each client I would have to retrieve its new id. Now I know that
> some databases like MS Access don't let you specify a value for an
> autonumber field, and in most cases there's no reason to specify it,
> but I think that the use case is enough to question the logic of
> always assuming that if the id is >0 then it's not a new record, if
> there is a more straightforward and accurate way to determine that.
>
> On Tue, Aug 25, 2009 at 11:17 AM, David
> Pollak<[email protected]> wrote:
> >
> >
> > On Tue, Aug 25, 2009 at 7:08 AM, Naftoli Gugenheim <[email protected]
> >
> > wrote:
> >>
> >> I had seen that thread, but it doesn't make sense to add all that code
> to
> >> all of my mapper models when they're ordinary mappers with ordinary
> >> auto-increment primary keys, which I happen to be importing from the
> >> database used by the previous, PHP version of this app. The foreign keys
> >> have to match up properly.
> >> Meanwhile I changed it so that it keeps a Map in memory from old ids to
> >> mapper instances, and linked children to the mapper with the new id.
> >> But wouldn't the logic of saved_? make more sense the way I proposed?
> >
> > For the load operation where you're moving databases, why not just write
> > some custom code to dump the data?
> >
> >>
> >> -------------------------------------
> >> Peter Robinett<[email protected]> wrote:
> >>
> >>
> >> You need to set the mapper id field as editable.
> >>
> >> Here is what I had to do to use a MappedStringIndex as a primary key
> >> that I can set manually:
> >> /* MAC address as primary key */
> >> def primaryKeyField = mac
> >> object mac extends MappedStringIndex(this, 17) with IndexedField
> >> [String] {
> >> override def dbDisplay_? = true
> >> override lazy val defaultValue = randomString(maxLen)
> >> /* allow user-defined primary key */
> >> override def writePermission_? = true
> >> override def dbAutogenerated_? = false
> >> private var myDirty = false
> >> override def dirty_? = myDirty
> >> override def dirty_?(b : Boolean) = { myDirty = b;
> super.dirty_?(b)
> >> }
> >> override def fieldCreatorString(dbType: DriverType, colName:
> >> String):
> >> String = colName+" CHAR("+maxLen+") NOT NULL "
> >> }
> >>
> >> Peter
> >>
> >> On Aug 24, 5:30 pm, Naftoli Gugenheim <[email protected]> wrote:
> >> > I'm trying to import data from a non-mapper-based database to one that
> >> > is. I'm reading data with DB.exec and trying to save it using Mapper,
> but
> >> > it's not getting saved. I'm guessing that since I'm setting the id (in
> >> > runSafe) it thinks it needs to do an update, not an insert. Is there
> a
> >> > workaround? Also, doesn't this scenario break the logic of
> >> > MetaMapper.saved_?() ? Wouldn't it make more sense to have a saved
> flag that
> >> > is cleared by default and set when it's created as a result of a find,
> and
> >> > after it's saved?
> >>
> >>
> >>
> >
> >
> >
> > --
> > Lift, the simply functional web framework http://liftweb.net
> > Beginning Scala http://www.apress.com/book/view/1430219890
> > Follow me: http://twitter.com/dpp
> > Git some: http://github.com/dpp
> >
> > >
> >
>
> >
>
--
Lift, the simply functional web framework http://liftweb.net
Beginning Scala http://www.apress.com/book/view/1430219890
Follow me: http://twitter.com/dpp
Git some: http://github.com/dpp
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Lift" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/liftweb?hl=en
-~----------~----~----~----~------~----~------~--~---