Release Date
April 13, 2005
Functional Changes
-
New Data Mining Edition - This is the first release of S4PM with a functional Data Mining Edition.
-
New Handling for Fake Export - Preventing some data types from being exported has gotten a lot
easier in this release. In previous releases if you wanted to
prevent some data type from getting exported, you had to configure
S4PM to cause the data to fail QC checks (e.g. by setting the
nominal output size to something unreasonable). In this release,
individual data types can be configured to go into the bit bucket
in the Export Station.
Detailed File Changes
s4p_import.pl
- New script to support the Data Mining Edition of S4PM in the
Stage For Pickup station. The script takes PDRs from the
Export station and ftps the files listed to a directory on
another box for distribution. Upon success, a PAN is sent
back to Receive PAN station.
s4pm_configurator.pl
- Fixed so that it detects whether the current string was
built using Stringmaker or Stringmaster and behaves accordingly.
s4pm_derived.cfg
-
A number of changes were put in to support Data Mining
Edition (when $dme is set). The link to
s4pm_allocate_disk.cfg in the Select Data station was
corrected, setting it to an absolute path rather than a
relative one (since Select Data is in an automounted
location).
-
Added in code for the new Stage For Pickup
station. Modified the Export station so that PDRs are sent
to the Stage For Pickup station. Modified the Receive PAN
station so that it looks for matching PDRs in the new
location that Stage For Pickup
writes to.
-
Another fix so that the s4pm_regular_block.pl
script is set up as a link in the Register Data station
(it's already there in the Register Local Data station).
-
Also in Register Data, the same %trigger_block hash that's
written to the s4pm_register_data.cfg file in Register Local
Data is now written to Register Data as well. We can do this
since NEWDATA work orders go to either Register Data or
Register Local Data, but not both.
-
Modified the command line using s4pm_run_easy.pl so that it
includes the -l and -v options.
-
Made a fix to the %cfg_commands hash in the Split Services
station that prevented large orders from being split up (TT
14552).
s4pm_export.pl
-
Updated to handle ersatz export (sending data into the bit
bucket) as it already handles export to datapool and
archive. The change allows individual data types to be
marked for ersatz, datapool, or archive within the same string.
s4pm_preselect_data.pl
-
Fixed the way the file accumulation production rule is
handled. The code for handling this rule was rigged so that
it would only work for accumulation periods that were one
day. New code was added so that the accumulation period
could be minutes, hours, days, or weeks. The only limitation
is that it assumes that the accumulation period is aligned
to the minute, hour, day, or week boundary. Thus for
example, you cannot have a 24 hour accumulation period that
starts 12 hours before the day and goes until 12 hours after.
s4pm_receive_pan.pl
s4pm_select_data.cfg
s4pm_stringmaker_algorithm.cfg
s4pm_stringmaker_derived.cfg
-
Minor fix so that when $dme is enabled, the link to
s4pm_allocate_disk.cfg in the Select Data station is to an
absolute path rather than a relative one (since Select Data
is in an automounted location).
-
Another fix so that the s4pm_regular_block.pl script is set
up as a link in the Register Data station (it's already
there in the Register Local Data station).
-
Also in Register Data, the same %trigger_block hash that's
written to the s4pm_register_data.cfg file in Register Local
Data is now written to Register Data as well. We can do this
since NEWDATA work orders go to either Register Data or
Register Local Data, but not both.
-
Modified the command line using s4pm_run_easy.pl so that it
includes the -l option.
-
Made a fix to the %cfg_commands hash in the Split Services
station that prevented large orders from being split up (TT
14552).
|
 |