Hi David,
When you say you want to store the update history, do you mean you want to
store just when the document was updated or do you mean you actually want to
store the change deltas?
I think properties will work either way, but if you just need to store when the
doc was updated for each update, that is probably pretty easy to do (and if you
just needed to answer "all documents updated since XXX" then the last-modified
system-updated property may even be enough).
If you store the updates as some datetime format and then put a range index on
that property (or put a range index on the prop:last-modified element if the
last-updated is good enough), then a simple range-query search against
properties will give you what you want I think.
Off the top of my head, you can put some xml like this in a property (using
xdmp:document-add-properties):
<updated>2010-04-15-07:00</updated>
Have an xs:date range index on the "updated" element.
then you can search it something like the following, which returns the URIs of
every document updated after April 15 2009:
for $x in cts:search(xdmp:document-properties(),
cts:element-range-query(
xs:QName("updated"), ">",
xs:date("2009-04-15-07:00") ) )
return
xdmp:node-uri($x)
Now I have not tried this so I am not sure how well it would work, but it seems
reasonable to me.
-Danny
From: [email protected]
[mailto:[email protected]] On Behalf Of Lee, David
Sent: Thursday, April 15, 2010 9:17 AM
To: General Mark Logic Developer Discussion
Subject: [MarkLogic Dev General] Update Logs ...
An upcoming project I need to maintain a log of all updates to documents. The
MarkLogic DB is being used to 'mirror' a dataset with changes occuring daily.
About 20,000 documents but expected about 10 change daily. For this were not
using the Library API (may at a future point but not now).
I was planning on using document properties to store update history. I still
think this is a good idea, but there's the need to produce a global report of
"all documents updated since XXX". Would this work well by simply querying
the document properties ?
Another Idea I had is keeping a directory and adding small 'log files' to it
for every update that contain the URI and update date.
Much like I'd do in an RDB (add a new record for every activity).
I suspect this would be in a directory and could grow large over time (but not
that large as the change rate is low). However to do this I need to create
unique ID's. Is this any better then using the document properties ?
Alternatively I could simply append elements to a single audit XML document ...
but I suspect that requires loading, inserting, then storeing the document
every time and it would grow unbounded.
I'm sure this is a common problem, any suggestions ?
----------------------------------------
David A. Lee
Senior Principal Software Engineer
Epocrates, Inc.
[email protected]
812-482-5224
_______________________________________________
General mailing list
[email protected]
http://xqzone.com/mailman/listinfo/general
