Level 2

Solved

Query getNodes() Extremely Slow

9 replies
2072 views
Translate All

I am building a query using QueryBuilder and the query returns 32k results. I am using the following to create an Iterator so that I can index over each result.

Iterator<Node> itr =  query.getResult().getNodes();

This line however takes a really long time and is slowing my page down drastically. How can I speed this up?

This post is no longer active and is closed to new replies. Need help? Start a new post to ask your question.

joerghohAccepted solution

Adobe Employee

In Jackrabbit and Oak the complete execution time of a query always consists of the exucution of the query itself and the retrieval of the results. This is because it's quite common that you're query returns lot of results, but you never retrieve them all, so it is split. And 32k results is quite a number.

I am not in the details, but for me it looks like that this command (especially the query.getResult() part) takes that long because it fetches a good number of results from the repository. And this means not only only loading nodes, but there are also a lot of other acitivites (e.g. validating ACLs, so you only the pages you have at least read access to, etc.).

kind regards,
Jörg

Peter_Puzanovs

Community Advisor

QueryBuilder is great fun to use and improve :)

Does your query contains search for custom fields you have created ?

When you go inside the /system/console/jmx/com.adobe.granite%3Atype%3DQueryStat do you see your query there ?

Thanks,

Peter

Translate

joerghoh

Adobe Employee

When you use "order by", the complete result set must be loaded from the repo in order to get sorted properly. Try to avoid ordering if you don't need it, it can speed up the performance significantly.

kind regards,
Jörg

Translate

O

ogill

Adobe Employee

Can you share the query with us?

Translate

J

joshua103usAuthor

Level 2

Jörg Hoh wrote...

In Jackrabbit and Oak the complete execution time of a query always consists of the exucution of the query itself and the retrieval of the results. This is because it's quite common that you're query returns lot of results, but you never retrieve them all, so it is split. And 32k results is quite a number.

I am not in the details, but for me it looks like that this command (especially the query.getResult() part) takes that long because it fetches a good number of results from the repository. And this means not only only loading nodes, but there are also a lot of other acitivites (e.g. validating ACLs, so you only the pages you have at least read access to, etc.).

kind regards, Jörg

Then this means there is nothing I can do to increase the speed of this function?

Translate

J

joshua103usAuthor

Level 2

Pacoolsky wrote...

what is the exact query ? does it contain order by clause ?

Yes, it does Order By

Translate

J

joshua103usAuthor

Level 2

So, there is really no way to speed this process up?

Translate

joerghoh

Accepted solution

Adobe Employee

In Jackrabbit and Oak the complete execution time of a query always consists of the exucution of the query itself and the retrieval of the results. This is because it's quite common that you're query returns lot of results, but you never retrieve them all, so it is split. And 32k results is quite a number.

I am not in the details, but for me it looks like that this command (especially the query.getResult() part) takes that long because it fetches a good number of results from the repository. And this means not only only loading nodes, but there are also a lot of other acitivites (e.g. validating ACLs, so you only the pages you have at least read access to, etc.).

kind regards,
Jörg

Translate