Contents
There are important issues and limitations
and considerations of SQL Server when it comes to objects, indexes, data
access, locking, fields, and join considerations. Please refer to the file Differences_DG400_and_DSS.htm for
a listing of the Differences between Acceler8DB, DataGate for DB2/400 (DG400)
and DataGate for SQL Server (DSS).
When creating a physical
file on the iSeries using the database manager, it uses the system default for
the maximum number of records of 10,000. This can be very annoying
because each file has to then be changed manually (CHGPF) to accept more
records.
To fix this, DG/400 has been changed to create physical files with the "size(*nomax)" parameter of
the CRTPF command.
ADB 4.7 and ADB 5.0 cannot co-exist on NT
Server
Even if you install ADB
5.0 into a different folder than ADB 4.7, there is only one service (adbntsvc). The last
version installed will install the service and that service will point to the
version you installed last.
Backup your
Databases!!!
As with any other
information you keep on a PC, you should make sure that you keep 'current'
backups of your data. If you have ever
run a program like CHKDISK you will notice that
sometime 'clusters' get damaged on FAT file systems. Keeping a good backup outside of your PC will
enable you to recover from a disk failure.
Here
is a list of the main new features found in Acceler8DB release 5.0:
SQL
Server Support
With
ADB's SQL Server support, access to SQL Server is
provided via standard RPG file I/O operations such as CHAIN, SETLL and READ.
Through
Acceler8DB, you can use such features as QTEMP, libraries, Open Query File and
native iSeries 400 data types (such as packed) with SQL Server. You can access existing SQL databases, or
create new SQL Logical databases.
Database
Connection Pooling
When
database connection pooling is enabled, the communication to
the server and to the server resources are kept in a special ‘pool’. The next time the application requires
database services, one of the available connections in the ‘pool’ is assigned
to it. Thus, enabling
Database Connection Pooling results in faster database connections.
Multi-Threaded
Environment
Each
‘thread’ of every process gets a separate connection to the server. When the application closes the database, the
connection goes away, along with its corresponding support on the server.
Null
Field, Unicode Field and Variable Length Field Support
This
release allows the support of null fields, Unicode fields, and variable length
field support.
Copying
and Renaming DDS Files
You
can now copy and rename DDS files from the QTEMP library. This is done by selecting the DDS tab within
Tools – Options. When a file is copied
and created on an AS/400, a Data Description Specification file (DDS) is
created in a library called QTEMP.
·
Avoid
key buffer overflow when key size is near the "user limit" (dgMaxKeyLength) by allocating a large enough buffer for the
"internal limit" (dgMaxKeyBufferLength).
·
Fixed
printfile formatting character bug.
·
Remove
buggy assert() preventing proper range operations
against multiformat files. Effects
all ADB server platforms.
·
Allow
"delete file" operation to ignore "object not found"
condition when checking for (and if found, subsequently deleting) file def
cache entries. Provide for "recursive" range-mode calls on the
server.
·
Use
full 32 bits for byte offset file position.
·
Fix
coding error in template parameter specification (fixes archive creation bug).
·
Workaround
IBM bug in null field "write" processing (or its documentation).
·
Explicitly
specify CCSID of user for adbtask job, rather than
rely on override-able defaults.
·
Fix
for coding error in creating DDS definition for DBCS fields.
·
Derive
correct fieldtype of concat
field using base field def info instead of unreliable primary field def info.
·
Fix
cache synchronization logic bug in readSeqEqual().
·
Trap
memory exceptions which may indicate a parameter mismatch between caller and callee (CALL PARM).
·
Use
new null value status value as an enumerated value, not a bit value.
·
Use
task instance storage for errors *received* by as400 server, such as those
received by Datagate from the license manager (this
was lost in the DGMTSERVER changeover).
·
Only
set null key map bit (for keyed access) when the key has a null-capable part.
·
Fixed
another misinterpretation of db2 join file definition.
·
Accommodate
4.7 client multiformat access
in new network blocking code.
·
Use
given keylength for partial key comparisons.
·
Don't
clobber exception in readSeqPartKey().
·
Check
for mts/com+ "ObjectContext",
which we will presume indicates a middle-tier presence, and thus restrict adbcom to the terms of the MTS license.
·
Fix
logic error in select/omit expression handling.
·
In
Work with Libl, changed only the portion of the LIBL
that the user touched.
·
Preserve
Join Default setting when doing OpenQuery.
·
The
field "Text" is now used for the grid column heading instead of
"Column Headings". If
"Text" is blank, then the first non-blank "Column Heading"
is used.
·
Fixed
problems with opnqryfile.
AS/400 Only:
·
Determine
if the object for which a cache member is built resides in Qtemp. If so, add the object to a list of object to
be deleted at the end of the job. Delete cache members corresponding to a file
to be deleted.
·
Create
select/omit expressions for join files.
·
Correctly
interpret timestamp field metadata.
·
Scan
query string for "<>" and "!=",
and replace with "¬=".
·
Use
the dgNetFormatAttr to provide "levelcheck" functionality.
· &