The recommendation from SAP is that SAP HANA should have three environments: DEV, QAS, PRD.
In consideration to the high cost of the HANA Box, it is feasible to use only two boxes:
Box_1 (DEV and QAS)
DEV_schema : replicated tables from ECC-DEV
DEV_package : HANA objects like At_Views, An_views (that reads data from DEV_schema tables) and Ca_Views.
QAS_schema : replicated tables from ECC-QAS
QAS_package : HANA objects like At_Views, An_views (that reads data from QAS_schema tables) and Ca_Views.
Box_2 (PRD)
PRD_schema : replicated tables from ECC-PRD
PRD_package : HANA objects like At_Views, An_Views (that reads data from PRD_schema tables) and Ca_Views
SLT_1: reads from ECC-DEV and ECC-QAS. writes data into BOX_1, schemas DEV and QAS respectively.
SLT_2: reads from ECC-PRD. writes data into BOX_2, PRD schema.
Is this correct understanding from what SAP is recommending?
A few problems that arises with this approach are:
Moving from DEV to QAS is a manual process:
Step 1: export XML file from DEV_PACKAGE objects like An_Views and At_Views
Step 2: replace "DEV_ SCHEMA" for "QAS_SCEMA" in the XML file as it is a TXT file.
Step 3: import updated XML file in the QAS_PACKAGE, and activate objects
When moving procedures or SQLScript calculations views from DEV to QAS, manual corrections need to be done in the code at
- Package hierarchy when reading data from columns views
E.g.: lt_centr = CE_JOIN_VIEW("_SYS_BIC"."p.dev.0/AT_CENTR") ;
in this sample, package hierarchy "p.dev.0" is fixed and needs to be updated manualy to "p.qas.0".
- Schema name when reading data from tables
E.g.: lt_tax = SELECT DISTINCT "MANDT", "MWSKZ" FROM "DEV_SCHEMA"."J_1BAJ"
in this sample, "DEV_SCHEMA" is fixed and needs to be updated manually to "QAS_SCHEMA".
Please, feel free to comment/suggest/ask about this approach we are using in our project.
Regards,
Mauricio Cubillos