(Re: 32-bit vs 64-bit) Of the two machines I've been trying this on, one is my laptop, and it turned out that it was using 32-bit Java. So I upgraded to the 64-bit version, and I find that I can actually get it to work, more or less, with Xmx4096m. So thanks for that. The other machine has had 64-bit Java all along, and I still get an out of memory error even with Xmx4096m, but it's a VM so there might be resource-sharing issues - I'll come back to that later.
So, about running this on the command line, here's the current state of the code:
let $map := map:merge( for $part in //part return map:entry(string-join($part/partinfo/*, '/'), $part) ) for $partinfo in //unit/partinfo let $part := $map(string-join($partinfo/*, '/')) return replace node $partinfo with <part>{$part/node()}</part> for $page in //page return file:write('data/'|| $page/@key || '.xml', $page, map{"omit-xml-declaration":"no"})
1) There'll be a new version of the input file arriving periodically, so I'd prefer to be able to do everything at the command line without having to create the database inside the GUI. Not sure of the best way to go about that.
2) In the GUI for now, I can perform the replace with the map, or I can run the for loop that writes out all the files. But I get an "Unexpected end of query" error when I try to do both, so what's the problem with my syntax?
Thanks so much,
Michael
On Wed, May 25, 2016 at 11:09 PM, Christian Grün christian.gruen@gmail.com wrote:
Hi Michael,
Now for basex.bat, in order to create a context, I started the script
with
'declare context item := doc("input.xml");' which may not be the most efficient way to do this, I don't know.
If you have created a database, you can use the command-line flag -i:
basex.bat -i input query.xq
The db:open function can be used as well:
declare context item := db:open("input");
Is it only the query with the map constructor that requires more memory?
Best, Christian
But on the command line or in the
GUI, I haven't had any luck.
Any other suggestions?
Thanks,
Michael
On Tue, May 24, 2016 at 10:49 PM, Christian Grün <
christian.gruen@gmail.com>
wrote:
Usually, 8GB should be much more than sufficient for such a query. You could try to increase the memory, which is assigned to Java, in the start scripts [1].
Does this help? Christian
[1] http://docs.basex.org/wiki/Start_Scripts
On Tue, May 24, 2016 at 11:52 PM, Michael Sanborn galethog@gmail.com wrote:
Seems like this would be perfect. I do need both number and manuf.
Using
your combination map, I'm now getting an "Out of Main Memory" error. Tried on a second computer - same issue. Would it be more likely to work if
I
tried it from the command line rather than the GUI? If so, I'll need
to
look up how to create a database that way, but I'm sure it's close to hand. Or is there a better workaround (besides buying a computer with more than
8GB
of RAM)?
Thanks again,
Michael
On Tue, May 24, 2016 at 2:10 PM, Christian Grün christian.gruen@gmail.com wrote:
Maybe you need something like this:
for $partinfo in //unit/partinfo for $part in //part[deep-equal(partinfo, $partinfo)] return replace node $partinfo with $part/node()
The deep-equal will be pretty slow. If the value of the number
element
is unique, you could do something like this:
for $partinfo in //unit/partinfo let $number := $partinfo/number let $part := //part[partinfo/number, $number] return replace node $partinfo with $part/node()
Using a map will even be faster:
let $map := map:merge(//part/map:entry(partinfo/number/text(), .)) for $partinfo in //unit/partinfo let $part := $map($partinfo/number) return replace node $partinfo with $part/node()
If you need to consider both number and manuf, you could e.g. combine these two in the map:
let $map := map:merge( for $part in //part return map:entry(string-join($part/partinfo/*, '/'), $part) ) for $partinfo in //unit/partinfo let $part := $map(string-join($partinfo/*, '/')) return replace node $partinfo with $part/node()
Does this help? Christian
On Tue, May 24, 2016 at 10:54 PM, Michael Sanborn <
galethog@gmail.com>
wrote:
Thanks for that. The trouble in step 2 is, just wrapping partinfo with the part element doesn't get me what I've labelled "misc part content
1"
and "misc part content 2". It's not sufficient to have just the tags -
I
need all the content of the corresponding part elements in the later
part
of the file. Is that something that can be done without too much
difficulty?
Thanks,
Michael
On Tue, May 24, 2016 at 12:16 PM, Christian Grün christian.gruen@gmail.com wrote: > > Hi Michael, > > Yes, this can easily be done with XQuery. There are many ways to
do
> this; here is just one: > > 1. First, create a database from your input file (e.g. with the > BaseX > GUI) > > 2. Second, run the following query to replace wrap your partinfo > elements with part elements: > > //unit/partinfo/(replace node . with <part>{ . }</part>) > > 3. Third, write all page elements to disk: > > for $page at $c in //page > return file:write($c || '.xml', $page) > > Hope this helps, > Christian > > > > On Tue, May 24, 2016 at 8:54 PM, Michael Sanborn > galethog@gmail.com > wrote: > > I need to perform a transformation that would be simple in XSLT, > > but > > the > > input is a file about 250 MBs in size. I'm wondering whether > > XQuery > > and > > BaseX in particular would be the most efficient way of doing it. > > I'm > > new > > to > > XQuery, and I've come up with a couple of ways to do this, but > > they > > turn > > out > > to be very time-consuming, so I'm sure I'm Doing It Wrong.
Hoping
> > to > > find > > out the proper way of doing this. > > > > The input consists of 2 sections. There are about 3600 page > > elements > > with > > this structure: > > > > <page> > > [misc page content...] > > <list> > > <unit> > > [misc unit content 1...] > > <partinfo> > > <number>54321</number> > > <manuf>A321</manuf> > > </partinfo> > > <partinfo> > > <number>12345</number> > > <manuf>B123</manuf> > > </partinfo> > > [misc unit content 2...] > > </unit> > > [multiple units...] > > </list> > > </page> > > > > Each unit can have 1 or 2 partinfo elements. The other section
has
> > about > > 82000 part elements like this: > > > > <part> > > <partinfo> > > <number>54321</number> > > <manuf>A321</manuf> > > </partinfo> > > [misc part content 1] > > </part> > > [...] > > <part> > > <partinfo> > > <number>12345</number> > > <manuf>B123</manuf> > > </partinfo> > > [misc part content 2] > > </part> > > > > I want to replace each unit/partinfo with the correpsonding
part,
> > like > > this: > > > > <page> > > [misc page content...] > > <list> > > <unit> > > [misc unit content 1...] > > <part> > > <partinfo> > > <number>54321</number> > > <manuf>A321</manuf> > > </partinfo> > > [misc part content 1] > > </part> > > <part> > > <partinfo> > > <number>12345</number> > > <manuf>B123</manuf> > > </partinfo> > > [misc part content 2] > > </part> > > [misc unit content 2...] > > </unit> > > [multiple units...] > > </list> > > </page> > > > > Is BaseX a good tool for this task? If so, how does one go about > > it? > > > > Finally, it would help to be able to output each page element
in a
> > separate > > file. Would it be better to have BaseX do this, or to output the > > whole > > database and chunk it with another tool? > > > > Thanks, > > > > Michael