When the number of rows in a dataset is not know then server side pagination can still be achieved by requesting one more row than the page limit and looking if that extra row exists in the returned data (if it does not then we know we are at the last page).
The documentation says that irrespective of the requested number of rows of data from a data read request there is a hard limit
to what can be sent over the wire.
There is an edge case here where the client side UI does not know what the hard limit is and could accidentally identify the end page of a dataset.
For example if a dataset has 101 rows and a page size is 50 then there are three pages. If the hard server limit of returned records is say 1000 there is no problem. The UI can request records 1-51, 51-101 and 101-151 to work out the number of pages.
However if the hard limit is 50 then the UI will say that there is only 1 page (a request for rows 1-51 will return 1-50).
Potential Solutions:
Add the hard coded data limit to the platform information API endpoint or new secure endpoint (with things like the repository config in it) so that this edge case can be avoided.
Feature Request
Description of Problem:
When the number of rows in a dataset is not know then server side pagination can still be achieved by requesting one more row than the page limit and looking if that extra row exists in the returned data (if it does not then we know we are at the last page).
The documentation says that irrespective of the requested number of rows of data from a data read request there is a hard limit to what can be sent over the wire.
There is an edge case here where the client side UI does not know what the hard limit is and could accidentally identify the end page of a dataset.
For example if a dataset has 101 rows and a page size is 50 then there are three pages. If the hard server limit of returned records is say 1000 there is no problem. The UI can request records 1-51, 51-101 and 101-151 to work out the number of pages. However if the hard limit is 50 then the UI will say that there is only 1 page (a request for rows 1-51 will return 1-50).
Potential Solutions:
Add the hard coded data limit to the platform information API endpoint or new secure endpoint (with things like the repository config in it) so that this edge case can be avoided.