|
 Handling
large data. Although DataDiff can
handle (load, display and compare) pretty large amounts of data, it
is necessary to remember that all data from the two tables is
actually loaded
into memory. Depending on the amount
of RAM installed on your machine and other software running
simultaneously, this may be an additional heavy load on the system
resources which could slow down your system
significantly.
BLOB
threshold: The fastest way to
consume all your system resources is to compare tables with large
BLOB data in them. To avoid such situations, DataDiff would compute
a 16-byte MD5 "digest" for any BLOB value longer than the threshold
value specified in this dialog box. The MD5 digests are
almost
guaranteed to
be different if the original values were different, so they can be
compared instead of the original data. Of course, it is impossible
to recover the original data from the digest, so the digested
columns can not be displayed or scripted (unless you use
server-side scripting). By default, the cut-off size for BLOBs is
4Kb, but you can change it to anything up to 64Kb. The data storage
format in DataDiff buffers does not allow to keep values longer
than 64Kb each. Anyway, scripting large pieces of data as
0xXXXXX... literals is not be such a good idea; in this case
server-side scripting (see below) would be preferable.
Please note that in
recent Diff versions BLOB thresholds can be set separately for each
column of the compared tables. The threshold option works
just as a suggested default value in the
column configuration
dialog.
[x]
Don't compare BLOBs: (see snapshot).
Again, this is just a suggestion for the DataDiff that when you
start configuring a data comparison, all BLOB columns will be
initially marked as Ignored.
You will be
able
Server-side
scripting. When the two databases
are on the same server or the servers can talk to each other using
the "linked servers" mechanism, it is possible to synchronize data
between the databases without including the actual literal values
into the synchro scripts. Server-side scripting may be simply
convenient, or it may be the only way to synchronize data if the
values (a) very large (>64Kb) or (b) contain non-Latin Unicode
characters (AdeptSQL Diff 1.XX does not properly handle Unicode
strings, so characters that don't belong to either Latin or the
selected "default" character set may be turned into '????' on the
way).
Note that the server-side
options in this dialog only determine the default behavior.
When you configure a specific data
comparison, you can define (a) whether or not server-side
synchronization is possible at all and (b) individually set
synchronization modes for each compared column.
Display
whitespaces... If some of the compared
data values contain trailing spaces, it is sometimes difficult to
see why DataDiff marks them as changed. When this option is
enabled, any spaces are displayed as "middle dot" characters (e.g.
"Test·value····").
|
