Mamta Satoor wrote:
I spent some time on points 1(using Monitor to get dvd directly) and
3(optimized allocation, caching some of the work.) which requires us to
solve the problem of how to get the InstanceGetter to return the correct
DVD for character types. Let me first briefly describe how the
InstanceGetter works for DVDs currently.
I'm a little unclear on exactly the problem this is trying to solve. I
got a little lost in the details, but does it take account of the fact
that the registered format ids are system wide and there can be
databases with different default collations in the same system?
Also the use of InstanceGetters seem to complicate this issue, once one
knows one has a collation type and one is using the DataValueFactory
then one can have methods on DVF that return DataValueDescriptors
directly, no need to go through the indirection of InstanceGetters. They
are a mechanism used when the type of the object is not known, here the
type is known as a DVD.
One of the points to note is that the correct DVD type for collation is
only needed when collation is actually occurring. If a collator based
column is read in using SQLChar then it's not a problem as long as a
switch to the collator version occurs during comparisons. Earlier I has
suggested methods to perform this switch on StringDataType, something
like getCollationValue(int collationType).
Dan.
***********description on InstanceGetter for DVD********
I think the code dealing with getting an InstanceGetter for a DVD from a
formatid is currently isolated in BaseMonitor.classFromIdentifier(int
fmtId). BaseMonitor has a class level field called rc2 which is an array
of same length as StoredFormatIds.TwoByte. The elements in rc2 will be
InstanceGetters. Every time BaseMonitor.classFromIdentifier(int fmtId)
is called, the method first checks if there is already an InstanceGetter
in the rc2 array for the passed format id. If yes, then it simply
returns that cached InstanceGetter from rc2. But if this is the first
time this method is being called for the passed format id, then we first
get the name of the InstanceGetter from RegisteredFormatIds using the
format id passed to the method. (For DVDs, the name
of that InstanceGetter would be
org.apache.derby.iapi.types.DTSClassInfo). Using that name from
RegisteredFormatIds, we create a Class object(for DVDs, that Class
object would be DTSClassInfo) and check if that Class is of
type FormatableInstanceGetter. If yes, then we create an instance of
that Class object(for DVDs, this will return an object of type
DTSClassInfo) and set the format id on it. And as a last step, we cache
this FormatableInstanceGetter in the rc2 array for future. So, in
future, if BaseMonitor.classFromIdentifier(int fmtId) gets called for
the same fmtId, we can simply return the cached InstanceGetter from rc2.
************************************************************
This current code will work fine for non-character type DVDs in Derby
10.3 but it won't work for character type DVDs. For example for the
format id corresponding to SQL type CHAR, we want to return DVD of type
either SQLChar or CollatorSQLChar, depending on the value of collation
type. But existing code will always return SQLChar. What we want is for
one format id to represent 2 DVDs and the deciding factor is the
collation type. In order to support this, I am proposing following
changes to the logic above so that we can have InstanceGetter return the
correct DVD, even for character types.
**********************************changes proposed to
InstanceGetter******************
For collation sensitive format ids (those corresponding to character
types), I am proposing to create a new InstanceGetter class called
CollationSensitiveDTSClassInfo which will extend DTSClassInfo . We will
change RegisteredFormatIds.TwoByte for such format ids to use
org.apache.derby.iapi.types.CollationSensitiveDTSClassInfo. We will also
need to remove the code for collation sensitive format ids from
DTSClassInfo since they will be handled in the new InstanceGetter, which
is CollationSensitiveDTSClassInfo.This new InstanceGetter class will
have two additional fields called collatorForDVD and collationType. And
it will have 2 setter methods, namely, setRuleBasedCollator and
setCollationType. The public Object getNewInstance() method on this
InstanceGetter will have code like following (Note that, I will need to
add a new constructor on CollatorSQL.. classes to take just the
RuleBasedCollator.)
switch (fmtId) {
/* Wrappers */
case StoredFormatIds.SQL_CHAR_ID:
if (collationType == StringDataValue.UCS_BASIC)
return new SQLChar();
else
return new CollatorSQLChar(collatorForDVD);
case StoredFormatIds.SQL_VARCHAR_ID:
if (collationType == StringDataValue.UCS_BASIC )
return new SQLVarchar();
else
return new CollatorSQLVarchar(collatorForDVD);
case StoredFormatIds.SQL_LONGVARCHAR_ID:
if (collationType == StringDataValue.UCS_BASIC)
return new SQLLongvarchar();
else
return new
CollatorSQLLongvarchar(collatorForDVD);
case StoredFormatIds.SQL_CLOB_ID:
if (collationType == StringDataValue.UCS_BASIC)
return new SQLClob();
else
return new CollatorSQLClob(collatorForDVD);
default: return null;
}
The collatorForDVD will need to be set on this new InstanceGetter only
the first time around when it is created. If user has requested
territory based collation, then collatorForDVD will be set to the
Collator that is derived from the database's territory. If user wants
UCS_BASIC collation, then collatorForDVD will be set to JVM's default
Collator. The collationType is subject to change depending on if store
is looking for character types belonging to system tables (such types
will always have collation type of UCS_BASIC) or for character types
belonging to non-system tables (such types will have the collation type
of UCS_BASIC/TERRITORY_BASED depending on what user has requested for
the database). Based on this, the logic for
DVF.instanceGetterFromIdentifiers(fmtId, collationType) will look as follows
DVF will have a class level field called instanceGettersForFormatIds
which will be an array of same length as StoredFormatIds.TwoByte. The
elements in instanceGettersForFormatIds will be InstanceGetters. Every
time DVF.instanceGetterFromIdentifiers (int fmtId, int collationType)
will be called, the method will first check if there is already an
InstanceGetter in the instanceGettersForFormatIds array for the passed
format id. If yes, then it will check if the instanceGetter is of type
CollationSensitiveDTSClassInfo and if yes, then it will set the
collationType on that InstanceGetter to the collationType passed to
instanceGetterFromIdentifiers method and it will return that
InstanceGetter. If the InstanceGetter is not
CollationSensitiveDTSClassInfo, then it will simply return the
InstanceGetter obtained from the instanceGettersForFormatIds array.
In the case, DVF.instanceGetterFromIdentifiers(int fmtId, int
collationType) does not find InstanceGetter cached for the passed format
id in instanceGettersForFormatIds array, then it will first get the name
of the InstanceGetter from RegisteredFormatIds using the format id
passed to the method. (For non-character DVDs, the name
of that InstanceGetter would be
org.apache.derby.iapi.types.DTSClassInfo. For character DVDs, the name
of that InstanceGetter would be
org.apache.derby.iapi.types.CollationSensitiveDTSClassInfo). Using that
name from RegisteredFormatIds, we will create a Class object(for DVDs,
that Class object would be
DTSClassInfo/CollationSensitiveDTSClassInfo) and will check if that
Class is of type FormatableInstanceGetter. If yes, then we create an
instance of that Class object(for non-character DVDs, this will return
an object of type DTSClassInfo. For character DVDs, this will return an
object of type CollationSensitiveDTSClassInfo) and set the format id on
it. For non-character DVDs, as a last step, we will cache this
FormatableInstanceGetter in the instanceGettersForFormatIds array for
future. But for character DVDs, we will set the collationType and
RuleBasedCollator on the InstanceGetter AND then save it in
instanceGettersForFormatIds.
As usual, I might have provided lot of information but hopefully it will
help understand the logic clearly. I will start looking at implementing
this but if anyone has any feedback on the logic, I will appreciate that.
thanks,
Mamta
On 4/12/07, *Mike Matrigali* <[EMAIL PROTECTED]
<mailto:[EMAIL PROTECTED]>> wrote:
Mamta Satoor wrote:
> Mike, the following code will be part of DataValueFactory and
hence it
> will be part of the interface. Please let me know if I am not
very clear
> with what I am proposing or if you forsee problems with this logic.
> if (dvd instanceof StringDataValue)
> dvd = dvd.getValue(dvf.getCharacterCollator(type));
My comment isn't really the logic, I think we are just not talking about
the same area. I think the code above belongs hidden behind the new
interfaces in the implementation logic of the data factory and data
types, not an example of what callers of the datatype should be doing.
>
> Also, in the following line below
> "I'll look at building/using DataFactory interface. It will be some"
> you mean DataValueFactory interface, right?
>
> Mamta
Yes I meant DataValueFactory interface. Let's work together on getting
the DataValueFactory interface right.
So far I have uncovered to basic ways store creates "empty" objects.
Note that store really only needs "empty" objects, ie. it is going
to initialize the state of these objects from disk by calling each
objects readExternal() method. But we have decided to not store
the collation info as state in the object so somehow we need to get
that info into the empty objects.
The ways store currently creates these objects:
1) using Monitor to get dvd directly:
dvd = Monitor.newInstanceFromIdentifier (format id)
o I think this use is best implemented as Mamta suggests, just
providing a non-static interface on the DataValueFactory.
something like:
DataValueFactory dvf = somehow cache and pass this around store;
dvd = dvf.newInstance(format id, collation id);
at this point dvd can be used to correctly compare against other
dvd's in possible collate specific ways.
2) using existing dvd's class to get a new "empty" dvd that matches it
(which is why it does not call clone).
dvd = dvd.getClass().newInstance()
o less sure about this one. Seems like we need a new dvd interface
that does the equivalent thing. I believe the original code got
here because the original store code did not deal with DVD's it
just got objects, so could not make dvd calls. There is a
getNewNull() interface, anyone know if there is any runtime work
that would be saved over this by creating a
getNewEmpty() interface?
dvd = dvd.getNewEmpty();
at this point dvd can be used to correctly compare against other
dvd's in possible collate specific ways.
3) optimized allocation, caching some of the work. This is used
where one query may generate large number of rows - for instance
hash table scan and sorter calls. Here the idea is to do some
part of the work once leaving an InstanceGetter which then can
repeatedly give back new objects in the most optimized way:
called once:
InstanceGetter = Monitor.classFromIdentifier(format id)
called many times:
dvd = InstanceGetter.getNewInstance()
o something like the following would be the direct conversion. Note
that implementation of the Instance getter is probably more complex
now. It can't just remember a single class and call new instance
on it. It has to cache some info on what class to create and what
collation to set in it.
called once
DataValueFactory dvf = somehow cache and pass this around store;
InstanceGetter =
dvf.instanceGetterFromIdentifiers(format id, collation id)
called many times:
dvd = InstanceGetter.getNewInstance()
again at this point dvd can be used to correctly compare against other
dvd's in possible collate specific ways.
All 3 of these uses have to be replaced to allow store to create
"correct" types which can be used in possible string comparisons.
