General
Avoid executing set sync on before a get command as this is already the default.
Avoid executing set sync off after a get command and before executing the queries, as such queries will not trigger a sync (applicable for Fabric Version 5.x and above).
This is the default system behavior, but it can be modified via a change in the config.ini file.
On versions prior to 5.x, set sync off should be used.
Always attempt to use in-memory data and avoid accessing external interfaces.
Avoid accessing the same table multiple times if all queries can be combined into one query.
Stop an LUI sync when it takes too long by using the sync timeout property on the LU level.
Select references from within an enrichment function. Do this by using ludb.fetch or fabric.fetch, instead of FabricDB.fetch (fabric local interface defined in the implementation), as ludb.fetch or fabric.fetch are more efficient.
Use the clearThreadGlobals function to make sure thread globals are cleared at the end of the get process.
When possible, use the Fabric set output file command followed by a query execution instead of writing code. This will generate a CSV file that contains Fabric data.
LUDB Functions:
- The purpose of LUDB functions is to expand the SQL library.
- An LUDB function that does not have or does not use input parameters and is used within a query will be called multiple times and will return the exact same result for all rows.
Graphit:
Try using Graphit whenever possible and minimal (or no) java code to allow easier maintenance and readability. If you are not sure how to implement a specific functionality in Graphit, please contact the COE.
A Graphit file can be a standalone web service.
Use sql non-prepared as a node type only when it is needed, by default sql should be used, as it works much faster.
If you use get inside the Graphit file, make sure a redundant get is not executed in the web service.
Use the resources try or close to release the entry back into the GraphitPool. Without this step, Fabric will generate/compile a new Graphit file every time, which could negatively affect performance. Code Example (calling graphit from enrichment function): GraphitPool.Entry entry = getLuType().graphitPool().get("Customer360.graphit"); Graphit graphit = entry.get();** Object result = graphit.run();** entry.close();**
General
Avoid executing set sync on before a get command as this is already the default.
Avoid executing set sync off after a get command and before executing the queries, as such queries will not trigger a sync (applicable for Fabric Version 5.x and above).
This is the default system behavior, but it can be modified via a change in the config.ini file.
On versions prior to 5.x, set sync off should be used.
Always attempt to use in-memory data and avoid accessing external interfaces.
Avoid accessing the same table multiple times if all queries can be combined into one query.
Stop an LUI sync when it takes too long by using the sync timeout property on the LU level.
Select references from within an enrichment function. Do this by using ludb.fetch or fabric.fetch, instead of FabricDB.fetch (fabric local interface defined in the implementation), as ludb.fetch or fabric.fetch are more efficient.
Use the clearThreadGlobals function to make sure thread globals are cleared at the end of the get process.
When possible, use the Fabric set output file command followed by a query execution instead of writing code. This will generate a CSV file that contains Fabric data.
LUDB Functions:
- The purpose of LUDB functions is to expand the SQL library.
- An LUDB function that does not have or does not use input parameters and is used within a query will be called multiple times and will return the exact same result for all rows.
Graphit:
Try using Graphit whenever possible and minimal (or no) java code to allow easier maintenance and readability. If you are not sure how to implement a specific functionality in Graphit, please contact the COE.
A Graphit file can be a standalone web service.
Use sql non-prepared as a node type only when it is needed, by default sql should be used, as it works much faster.
If you use get inside the Graphit file, make sure a redundant get is not executed in the web service.
Use the resources try or close to release the entry back into the GraphitPool. Without this step, Fabric will generate/compile a new Graphit file every time, which could negatively affect performance. Code Example (calling graphit from enrichment function): GraphitPool.Entry entry = getLuType().graphitPool().get("Customer360.graphit"); Graphit graphit = entry.get();** Object result = graphit.run();** entry.close();**
