Counting, the second

Hello,

I want to write some servlets to access Tamino using Java API.
The application should supply only the query-hits, by adding requests interactive.
e.g.
search argument 1 = 2345 hits
search argument 2 = 5822 hits, linked with argument 1, total hits = 56.

I have tried the count function as a query, …accessor.query(“count(” +path +“)”).
It workes, but it’s very slow in nested schemas.
e.g.
count(Article/cml[molecule/@title~=“ethanol”]/fizPhysicalProperties/string[@title~=“property”][.~=“pressure”])')
count all documents, they include ‘ethanol’ as molecule and ‘pressure’ as property will take ~10 seconds at 1000 documents!

There is no difference, whether the nodes are indexed or not.

Under tamino 2.2, the cursor supplies the attribut ‘ino:count’, but this function doesn’t run under tamino 3.1
(set page size to 1, and then take the attribute ino:count, thats very fast, but incorrect)!

The new tamino API function ‘openCursor’ which is called by TStreamAccessor only supplies ‘ino:current’ and ‘ino:next’.

The contribution from ‘Scudman’ presumes a query over all.

I think, when tags are indexed, the count-information should be present.

I need an ADABAS/NATURAL command like ‘find number’ !!!

How can I get my query-hits faster? Somebody has an idea?


Thanks for reading!

Uwe Wanja

hello and sorry,

i have set the wrong index, respectively the wrong query.

I have indexing the elements or attributes with ‘standard-index’ and phrase the query with ‘~=’, or reverse, or i have forgotten, that standard-index is case-sensitive and all the rest of it, like a rookie.

so long
Uwe Wanja