Batch Setup Example
This section will guide you through an example of setting modules in Retail Home and syncing with POM. A basic implementation of MFP with forecasting in AIF might send all the following foundation data files (and their CTX files):
-
CALENDAR.csv
-
PRODUCT.csv
-
ORGANIZATION.csv
-
EXCH_RATE.csv
-
SALES.csv
-
INVENTORY.csv
-
RECEIPT.csv
-
TRANSFER.csv
-
MARKDOWN.csv
-
ADJUSTMENT.csv
-
ORDER_HEAD.csv
-
ORDER_DETAIL.csv
-
RTV.csv
-
RA_SRC_CURR_PARAM_G.dat
These files will be bundled into a single ZIP file named RAP_DATA.zip
. To configure the Customer Modules
for this batch implementation, perform the following steps:
-
At the top level, you may disable the RI module entirely (if it’s visible), because you are not implementing that solution
-
Enable and expand the RAP module. Within the
RAP_COMMON
sub-module, enable only the following components and disable the rest:-
RAP>RAP_COMMON>ZIP_FILES: RAP_DATA_ZIP
-
RAP>RAP_COMMON>SICONTROLFILES: RAP_SI_DIM_CALENDAR, RAP_SI_DIM_EXCHANGE_RATES, RAP_SI_DIM_ONORDER, RAP_SI_DIM_ORGANIZATION, RAP_SI_DIM_PRODUCT, RAP_SI_FACT_ADJUSTMENT, RAP_SI_FACT_INVENTORY, RAP_SI_FACT_MARKDOWN, RAP_SI_FACT_ORDER_DETAIL, RAP_SI_FACT_RECEIPT, RAP_SI_FACT_RTV, RAP_SI_FACT_SALES, RAP_SI_FACT_TRANSFER
-
RAP>RAP_COMMON>SIBATCH: RAP_SI_INVADJ, RAP_SI_INVPOS, RAP_SI_INVRECEIPT, RAP_SI_INVRTV, RAP_SI_INVTRANSFER, RAP_SI_MARKDOWN, RAP_SI_ONORDER, RAP_SI_PO, RAP_SI_REQUIRED, RAP_SI_SALES
-
RAP>RAP_COMMON>RDXBATCH: RAP_RDX_INVADJ, RAP_RDX_INVPOS, RAP_RDX_INVECEIPT, RAP_RDX_INVRTV, RAP_RDX_INVTRANSFER, RAP_RDX_MARKDOWN, RAP_RDX_PO, RAP_RDX_REQUIRED, RAP_RDX_SALES
-
RAP>RAP_COMMON>BATCH: RAP_INVADJ, RAP_INVPOS, RAP_INVRECEIPT, RAP_INVRTV, RAP_INVTRANSFER, RAP_MARKDOWN, RAP_PO, RAP_REQUIRED, RAP_SALES
-
-
To use AIF for forecasting, you will also need certain parts of the AI Foundation modules (named RSP in Retail Home). Enable the RSP root module, expand it, and enable the FCST module. All options under FCST should also be enabled. Disable all other RSP modules here.
-
If you are integrating with Merchandising using the RDE batch jobs, then you are likely not providing flat files for most things, and the setup process will be different. Refer to Integration with Merchandising to understand the batch setup process in more detail.
Once all changes are made, make sure to Save the updates using the button below the table. Only after saving the changes will they be available to sync with POM.
The next set of steps will be performed from POM:
-
Go to the Batch Administration screen for each schedule (Nightly tabs of AIF DATA and AIF APPS schedules) and disable ALL jobs in the Nightly schedule. By disabling all jobs, the next step will selectively enable just what is needed.
-
Click Sync with MDF in each schedule. Wait for the sync to complete (it will take some time and the screen will prevent any updates while syncing).
-
Review the list of jobs that are now enabled to verify that it is correct (for example, that the jobs relating to PRODUCT file loads are all enabled similarly to when you ran historical data loads).
-
Inter-schedule dependencies must also be enabled at this time from Batch Administration screen. You will need to locate the jobs having dependencies and turn them on within the job details screen. Update the following jobs in each schedule to enable any inter-schedule dependencies they have:
Schedule Job Name AIF APPS
RSE_WEEKLY_INPUT_FILES_START_JOB
RPASCE
MFP_POST_DATA_IMP_START_JOB
RPASCE
RPASCE_DATA_IMP_START_JOB
RPASCE
RPASCE_HIER_IMP_START_JOB
-
Go to the Batch Monitoring screen. If you have a schedule already open for a past business date, click Close Schedule. If you are already on the current business date then just click Restart Schedule and skip the next step.
-
Change the POM business date to the date you will be loading nightly batch data for (for example if your data files will have data for 2/25/2023 then that should also be the business date). Open a new schedule using the provided button.
-
Enable and schedule the batches to run from Schedule Administration. The inter-schedule dependencies will ensure downstream jobs are not run until the necessary dependencies are complete (for example, RSP jobs will wait for RI loads to complete).
Note:
Updating the schedule times also requires a Restart Schedule action to be performed afterwards. -
If you have not already enabled the schedule for the
RI_MAINTENANCE_ADHOC
flow, make sure to do that now. -
Make sure the business date in the data warehouse is one day prior to the first day of data you plan to load through the nightly batch. If necessary, use the standalone POM process
LOAD_CURRENT_BUSINESS_DATE_ADHOC
to advance the business date. For example, if your first nightly batch is loading data for2023-04-30
then you must ensure the data warehouse is currently on business date2023-04-29
. The nightly batch will automatically advance the date starting with2023-04-30
. -
If you chose not to provide the
RA_SRC_CURR_PARAM_G.dat
file for ensuring proper business date validation, then go back to Batch Administration in the AIF DATA nightly schedule and disableBATCH_VALIDATION_JOB
,RA_SRC_CURR_PARAM_G_COPY_JOB
, andRA_SRC_CURR_PARAM_G_STG_JOB
.
The next set of steps may or may not be needed depending on your business calendar configuration:
-
From the Batch Administration Nightly schedules, locate the column for Days of the Week. By default, weekly jobs may run on either Saturday or Sunday (varies by process). You must align the weekly jobs to your week-ending dates for anything involved in integration or calculations, such as RDX and PDS batch jobs.
-
For example, if you wish to set your week-ending date as Saturday for all jobs involved in RI > Planning integrations, filter the AIF DATA Schedule job names using each of the following codes: PDS, RDX
-
For each set of PDS or RDX jobs, look at the Days of the Week value. Anything that is not set to run Saturday should be modified by editing the job record and changing the day to Saturday. You might also want some jobs to run daily, in which case you select all days of the week here.
-
Repeat this process in the RPASCE schedule, moving any jobs to the desired Day of the Week value.
-
When all changes are done, be sure to Restart Schedule from the Batch Monitoring screen.
Lastly, ensure that the following jobs are disabled in the AIF DATA schedule after all other setup is done, as they should not be used in the current version:
-
OBIEE_CACHE_CLEAR_JOB
-
ODI_LOG_EXTRACTOR_JOB
-
ODI_LOG_LOADER_JOB
You are now ready to begin running your nightly batches. As soon as you run the first batch cycle, you must continue to run AIF DATA and AIF APPS batch cycles every night in sequence. You must not skip any days, because some jobs have internal processing based on the day of week that they execute and skipping days will prevent proper operation of these jobs. If you are not providing daily fact data, then you must still run the daily batches with a full set of dimension files, such as products and locations; you can never run a batch without dimension files present and populated with data. The recommended approach is for the source system providing data files to always package and upload a nightly ZIP file every day even when no data is changing. The uploaded ZIP should contain the dimension data plus empty fact files where daily data is not being generated.