So the objection is that content-specific bits, such as the timestamp check,
would have to be hard-coded? There is a generic version of conditional updates
in the REST API. It relies on ETag, as described at
http://docs.marklogic.com/guide/rest-dev/documents
> When content versioning is enabled, MarkLogic Server associates an opaque
> version id with a document. The version id changes each time you update the
> document through your REST API instance. The version id is returned in the
> ETag header when you read a document, and you can pass it back in an If-Match
> header during an update or delete operation to test for changes prior to
> commit.
Granted that isn't quite the same as a timestamp check. But you could build
your own mechanism with a similar level of generic functionality. You could
parameterize your timestamp check so that it can use any relative XPath
compatible with xdmp:unpath. The code might look something like this:
...
declare variable $OLD := doc($URI) ;
if (empty(doc($URI))
or $TIMECODE lt ($OLD ! xdmp:unpath($PATH))) then ()
else xdmp:document-insert(
$URI, xdmp:unquote($XML),
xdmp:default-permissions(), local:getCollections())
Except for $OLD, all the capitalized variables would come from the $params map.
Now there is nothing content-specific about the code, and you can use it with
any content that has a timestamp at some known path.
If you cannot trust the caller to set the PATH parameter correctly, create a
lookup table using arbitrary codes. Use the caller-supplied code to look up the
correct path, or throw an error if it is invalid. The code for that would be a
little longer, but should still perform well.
-- Mike
On 19 Nov 2012, at 08:51 , Jem Rayfield <[email protected]> wrote:
> Extending the REST API every time you want to do constraint controlled
> Document Create, Update, Deletes seems like a fair amount of extra work.
> If this is the approach then I guess I can look at it.
>
> I was just thinking that perhaps something along the lines of would be
> availible:
>
> /documents&query=XQuery&bind:x=X&bind:y=Y
>
> Update /documents using the adhoc XQuery with a set of variable bindings.
> Where these bindings could be the XML for the doc and params.
>
> I guess I need to look at extensions.
>
> Thanks for the pointer re fn:exists().
>
>
>
> -----Original Message-----
> From: [email protected] on behalf of Michael Blakeley
> Sent: Mon 19/11/2012 4:41 PM
> To: MarkLogic Developer Discussion
> Subject: Re: [MarkLogic Dev General] Marklogic REST API. [Multi-statement tx]
>
> Jem, why not extend the REST API to implement that as a single request? That
> should give you the benefits of using the REST API, without requiring
> multiple requests.
>
> http://docs.marklogic.com/guide/rest-dev/extensions discusses this feature
> and has an example that you could modify to suit your application.
>
> BTW in your sample code below I would probably use fn:exists() rather than
> xdmp:exists(). You are using that boolean to decide whether or not to update
> the database. Therefore you probably want ACID consistency guarantees, which
> xdmp:exists explicitly avoids in its subexpression because it always runs
> timestamped. Replacing it with fn:exists(doc($documentUri)) will read-lock
> any document at $documentUri, which guarantees consistency and avoids a
> potential race condition.
>
> -- Mike
>
> On 19 Nov 2012, at 07:59 , Jem Rayfield <[email protected]> wrote:
>
> > Hi,
> >
> > Would I be correct that the following type of XQuery should be translated
> > to a sequence of REST calls using a tx-id.
> > So, a single simple XQuery update such as:
> >
> > if (xdmp:exists(doc($documentUri)))
> > then (
> > if ($timeCode < xs:dateTime(doc($documentUri)/document/header/timeStamp) )
> > then ()
> > else (
> > xdmp:document-insert($documentUri, xdmp:unquote($xml),
> > xdmp:default-permissions(), local:getCollections())
> > )
> > )else (
> > xdmp:document-insert($documentUri, xdmp:unquote($xml),
> > xdmp:default-permissions(), local:getCollections())
> > )
> >
> > Would be translated to a number of REST calls:
> >
> > Get a transaction; /transactions
> > Invoke XQuery to get time stamp using the tx id; /query
> > Update Document using the tx id; /documents
> >
> > Or is there a better way of doing this using the current REST API spec?
> > There could be a lot of I/O etc for something quite simple.
> >
> > I may have missed something here so please do point me in the right
> > direction.
> >
> > Cheers
> > Jem
> >
> >
> >
> > http://www.bbc.co.uk
> > This e-mail (and any attachments) is confidential and may contain personal
> > views which are not the views of the BBC unless specifically stated.
> > If you have received it in error, please delete it from your system.
> > Do not use, copy or disclose the information in any way nor act in reliance
> > on it and notify the sender immediately.
> > Please note that the BBC monitors e-mails sent or received.
> > Further communication will signify your consent to this.
> > _______________________________________________
> > General mailing list
> > [email protected]
> > http://developer.marklogic.com/mailman/listinfo/general
>
> _______________________________________________
> General mailing list
> [email protected]
> http://developer.marklogic.com/mailman/listinfo/general
>
>
>
>
> http://www.bbc.co.uk
> This e-mail (and any attachments) is confidential and may contain personal
> views which are not the views of the BBC unless specifically stated.
> If you have received it in error, please delete it from your system.
> Do not use, copy or disclose the information in any way nor act in reliance
> on it and notify the sender immediately.
> Please note that the BBC monitors e-mails sent or received.
> Further communication will signify your consent to this.
> _______________________________________________
> General mailing list
> [email protected]
> http://developer.marklogic.com/mailman/listinfo/general
_______________________________________________
General mailing list
[email protected]
http://developer.marklogic.com/mailman/listinfo/general
