cancel
Showing results for 
Search instead for 
Did you mean: 

Disable Caching for specific items

Former Member
0 Kudos

hi, we have large number of records (inserts) loaded to hybris application as part of initial data setup. What we have observed from JDBC log is that during the inserts application makes a check against the PK before executing the insert query. I think the select query is run against the cache (item/query) to see whether the record is existing. From JDBC log, the select query takes time. We would like to avoid hybris application making the select query by disabling caching during the initial load time. Snippets of JDBC log SELECT * FROM users WHERE PK=?

Can you please suggest how we can disable the caching in hybris application? Appreciate any additional inputs.

Thanks in advance

Accepted Solutions (0)

Answers (2)

Answers (2)

Former Member
0 Kudos

Thanks Christian.

For security issues (personal identifiable information), application uses service class to read from source and insert directly to database through batch mode (saveAll). For this table, application has a unique index on the PK. In this case, can we handle only inserts without the select statements.

One finding is that, whenever we have less cache entries, data load happens faster, we used to clear cache (through hAC) at different point of time to see how much the general load time is for the same batch sizes. Also earlier runs whenever we don't clear the cache, we have observed that the response of the SELECT query is large as more records are inserted.

Thanks Mano,

former_member602476
Active Participant
0 Kudos

Hi Mano, there is no way to disable caching, but I don't think caching is at fault here. What is your impex file looking like and how many rows you have currently in your users table. Make sure you have an unique index defined on the PK column in the users table, if you do the select should be pretty fast. Additionally, if you know your impex data does not exist in the database, you could alter your impex header to use only Insert instead of Insert_Update, I think that may get rid of the select query.