-
Type:
Task
-
Status: Resolved
-
Priority:
Minor
-
Resolution: Fixed
-
Affects Version/s: None
-
Component/s: Nuxeo Drive
Default contribution sets pageSize and maxPageSize to 1000: as we don't use pagination when fetching children, getting only the current page, we need to be sure we fetch all the child documents in most of the cases, yet having a limit for obvious performance reasons.
See DocumentBackedFolderItem#getChildren:
public List<FileSystemItem> getChildren() throws ClientException { PageProviderService pageProviderService = Framework.getLocalService(PageProviderService.class); Map<String, Serializable> props = new HashMap<String, Serializable>(); props.put(CORE_SESSION_PROPERTY, (Serializable) getSession()); PageProvider<DocumentModel> childrenPageProvider = (PageProvider<DocumentModel>) pageProviderService.getPageProvider( FOLDER_ITEM_CHILDREN_PAGE_PROVIDER, null, null, 0L, props, docId); List<DocumentModel> dmChildren = childrenPageProvider.getCurrentPage(); List<FileSystemItem> children = new ArrayList<FileSystemItem>( dmChildren.size()); for (DocumentModel dmChild : dmChildren) { FileSystemItem child = getFileSystemItemAdapterService().getFileSystemItem( dmChild, this); if (child != null) { children.add(child); } } return children; }
And nuxeodrive-pageproviders-contrib.xml:
<coreQueryPageProvider name="FOLDER_ITEM_CHILDREN">
<pattern>
SELECT * FROM Document WHERE ecm:parentId = ?
AND ecm:isCheckedInVersion = 0
AND ecm:isProxy = 0
AND ecm:currentLifeCycleState != 'deleted'
AND ecm:mixinType != 'HiddenInNavigation'
</pattern>
<sort column="dc:created" ascending="true" />
<pageSize>1000</pageSize>
<maxPageSize>1000</maxPageSize>
<property name="maxResults">PAGE_SIZE</property>
</coreQueryPageProvider>
In case this limit is not enough, the PageProvider can be overridden, as suggested in https://jira.nuxeo.com/browse/SUPNXP-11989?focusedCommentId=201193&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-201193
Let's have a unit test to ensure that this overridden contribution is taken into account.
PS: ideally we could use batching from Drive and lower page size.
