create index for queries in format "//"

, ,
1

query 1: /div[@id=“5”]/note
with index created on attribute “id”, the processor will use the index

query 2: //div[@id=“5”]/note
with multipath indexed created on attribute “id”, the processor will use the index

query 3: /div[@id=“5”]//note
query 4: //div[@id=“5”]//note
no matter what kind of index created on “id”, no index is used.

how can I create index so that query 3 and 4 can use the index.

2

Hi

again usage of a multiPath standard index (available with Tamino 4.2) is the best answer.

Best regards
Uli

3

as you can see, I tried to use multipath index on attribute “id”, which case make the second query use the index. I also have been able to create a index on “Div” to make the 4th query to use the index on attribute “id”.

However, no matter what index I created, the 3rd query just don’t use the index on attribute “id”

4

Hi Eric

sorry - I missed that you are already using the multiPath index.
Could you pls provide the schema you are using?

Thanks
Uli

5

here is the schema I used.

I actually simplify a real schema I used to this one. You can still just define this schema
against a temp collection. I have created multipath index on “div”, “id”, and even “note”,
with this schema,
/div[@id]/note ----- use the index “id”
//div[@id]/note------ use the index “id”

//div[@id]//note and /div[@id]//note don’t use the index “id”

some how, in my real schema, //div[@id]//note works. anyway, how can we
make the two queries to use the index on “id”

the scenario I have is : many users will update the documents in a collection at the same time using queries in above format, if Tamino doesn’t use the index, then it will lock the whole
doctype when a user update a document, preventing others to update different document at the same time.

6

Hi

it looks like there is no attachement - I just try to attach a dummy doc for testing.

Regards
Uli
dummy.txt (11 Bytes)

7

so sorry, I forget to attach.
Notebook_Simple.tsd (2.19 KB)

8

Hi

why do you think that no index is being used?
If you are using _xql=ino:explain(…,“tree”) you’ll see, that indexes
are being used for both queries
//div[@id=“myId”]/note
and
//div[@id=“myId”]//note
if the doctype does already contain documents contributing to the indexes.
This restriction is related to the special nature of multiPath indexes.

Best regards
Uli

9

thanks for checking out that for me. However, I did have document in my doctype, and I also used the “explain” button in “inoxplorer” for the xquery I used.

//div[@id=“ab”]//note

and

/div[@id=“ab”]//note

don’t use index.

I posted 4 screenshot for you. 1.jpg and 2.jpg are the screen for the two queries not using index. Success.jpg is for the query using the index. And file.jpg is the file I have in under my doctype.

Thanks a lot.
screenshot.zip (266 KB)