Hi,

Look at http://vschart.com/compare/basex/vs/exist-db

If you want, you can add other comparisons

Cheers, Ben

Op 18-04-18 om 16:34 schreef Alexander Holupirek:

...

On 18. Apr 2018, at 15:39, Feargal Hogan feargal.hogan@gmail.com wrote:

Hi

Is anyone aware of any comparisons between baseX and Exist? I have some familiarity with Exist and I’d like o understand what are the benefits of each.

Thanks

Feargal

Both are, of course, excellent systems. Do you have something special in mind that you would like to compare? Besides, I'm not aware of a general feature comparison site or something like that.

Cheers, Alex

On 18 Apr 2018, at 21:12, Liam R. E. Quin liam@w3.org wrote:

On Wed, 2018-04-18 at 14:39 +0100, Feargal Hogan wrote:

...

Hi

Is anyone aware of any comparisons between baseX and Exist? I have some familiarity with Exist and I’d like o understand what are the benefits of each.

What really matters is suitability to task, though, and that will depend on what you're trying to do. And part of suitability to task is the support network - are other people doing similar thigns using eXist-db or BaseX?

Liam

Hmmm, havent seen anyone doing what I am looking to do.

Initially, I want to replace filesystem storage for about 12k xml files with queryable storage.

As we progress, I may want to batch update mutiple records contextually and/or enhance the xml based on regex patterns.

From the comparison chart that Ben referenced earlier I noticed that baseX doesn’t seem to actually load xml files into an xml database, is that right? So what does it do then? It creates a queryable indexed representation of the files? Is that right?

And what happens when a file is edited/updated?

Does baseX need to be 'told' that it has been updated, in order to add the new data to its indeices? Or does it know there has been an update and automatically reindex?

Thanks

Feargal

Hi Feargal,

Just my two cents, but to stress the fact what Christian is saying: BaseX is an XML database (albeit the clever marketing guys at BaseX now branded it as "BaseX Framework" with the new webpage ;-) ), so of course it actually loads XML files into the database itself.

I am wondering why you want this evaluation: 12k documents sounds like... not much. Are these documents particularly large? Otherwise I would simple start with BaseX and put them all into the database and query the data. If your documents are not particularly huge that should be reasonably fast and you can basically evaluate this in ten minutes for yourself.

Also, I would like to add that BaseX (hence: A framework) is also a powerful XQuery processor. So if you want to "enhancve the XML with regex patterns" it sound technically inferior and also it makes sad pandas cry :( Why you should not use regex to parse XML, you ask? I kindly refer you to this excellent SO answer: https://stackoverflow.com/a/1732454/1451599

Cheers Dirk

Senacor Technologies Aktiengesellschaft - Sitz: Eschborn - Amtsgericht Frankfurt am Main - Reg.-Nr.: HRB 105546 Vorstand: Matthias Tomann, Marcus Purzer - Aufsichtsratsvorsitzender: Daniel Grözinger

On 19. Apr 2018, at 23:06, Christian Grün christian.gruen@gmail.com wrote:

Hi Feargal,

...

I noticed that baseX doesn't seem to actually load xml files into an xml database, is that right?

If you have a larger number of XML documents, and if the documents need to be processed multiple times, you will usually store them in a database. But it's generally possible with BaseX to process files without storing them in a database. But I would assume that this is possible with eXist-db as well.

I don't know who is maintaining the vschart.com web site, but I was wondering which information was misleading?

...

And what happens when a file is edited/updated?

Do you refer to the original file or a document in the database? If the original file is updated, it will need to be readded to your database.

...

Does baseX need to be 'told' that it has been updated, in order to add the new data to its indeices? Or does it know there has been an update and automatically reindex?

For more information on indexes in BaseX, I invite you to visit the corresponding article in our documentation [1], in particular the section on updates.

...

Thanks

Welcome, Christian

[1] http://docs.basex.org/wiki/Indexes

On Thu, 2018-04-19 at 16:26 +0100, Feargal Hogan wrote:

...

From the comparison chart that Ben referenced earlier I noticed that baseX doesn’t seem to actually load xml files into an xml database, is that right?

No. Yes. Maybe.

baseX does load the documents into a database. It stores them in an internal data structure, not as textual XML.

It creates a queryable indexed representation of the files? Is that right?

Yes.

And what happens when a file is edited/updated?

A file outside the database? Nothing.

Depending on database options, though, if you update a document in the db, the index is updated.

Does baseX need to be 'told' that it has been updated, in order to add the new data to its indeices? Or does it know there has been an update and automatically reindex?

This isn't a meaningful quesiton.

If you load a CSV file into a database such as Oracle, what happens if the CSV file changes on disk outside Oracle? And why do you care? You would normally edit the data at that point in the database using a SQL application.

BaseX doesn't need to consult the external XML files once the database is built (although yes, you _can_ keep files on disk and refer to them if you want, but then you're somewhat fighting the system and will have to go through some hoops to have super-fast queries).

As Christian and Dirk said, go give BaseX a try as many of your questions will be answered in some number of nanoseconds :) In particular, you can create a database from the GUI -- 12,000 files may take a few seconds to index, depending on how large they are -- and run queries directly.

One note on BaseX wth documents - it has an option to delete whitespace nodes on import, which, inappropriately for documents, is enabled by default. You'll find it in the Options tab when you make a database from the GUI, for example.

Liam

Hi Omar,

Thank you (and everyone else) for sharing your experiences.

• BaseX in comparison to exist-db turned out to be particularly bad at

hosting multiple XQuery based applications like RestXQ endpoint in one instance.

Definitely true; BaseX was not built for that. If you want to run multiple applications with a single web server, the recommended approach is to use the WAR distributions of BaseX and deploy each application as a separate servlet.

• BaseX is not as smart on recognizing when indices can be used in longer

XQuery code.

This one is interesting to hear, because we observed that users chose BaseX in the past exactly because of the index rewritings. Did you encounter these restrictions when working with multiple databases, or also with single instances?

Cheers, Christian

I use BaseX'es jobs for this. Works but you have to be careful because you give up all the protections against dead locking or becoming unresponsive because no new jobs can be scheduled anymore so all HTTP communication stops like database administration for example.

Best regards

Omar

Am 20.04.2018 um 15:57 schrieb Marco Lettere:

On 20/04/2018 15:49, Christian Grün wrote:

...

Hi Omar,

Thank you (and everyone else) for sharing your experiences.

...

• BaseX in comparison to exist-db turned out to be particularly bad at

hosting multiple XQuery based applications like RestXQ endpoint in one instance.

Definitely true; BaseX was not built for that. If you want to run multiple applications with a single web server, the recommended approach is to use the WAR distributions of BaseX and deploy each application as a separate servlet.

Or use it as we do here. A new BaseX process for every application. A sort of very small application container or nano-service with the optimization that they all share the BaseX code and just use different restxq and possibly data folders.

;-)

M.

Hi Omar,

Right, we had some discussion on this. If I remember correctly, you had been dynamically addressing databases in BaseX, whereas in eXist-db (as far as I remember), all queries and rewritten path expressions will refer to one static and globally opened database instance, right?

I guess your setup doesn’t allow this, but in use cases where the names of the databases are statically known, a popular approach is to reference the databases in global variables and address them inside the query. Here is a simple example for a nested query that will take advantage of the index:

declare variable $DB := db:open('factbook'); declare function local:city($city) { local:countries()//city[name = $city] }; declare function local:countries() { $DB//country }; local:city('Rome')

You can also open a global database (e.g. via basexhttp and the -c option [1]) and enable DEFAULTDB [2], but in this case, you’ll always need to use collection() or doc() before each path expression:

declare function local:city($city) { collection()//city[name = $city] }; declare function local:countries() { collection()//country }; local:city('Rome')

Hope this helps, Christian

[1] http://docs.basex.org/wiki/Command-Line_Options#HTTP_Server [2] http://docs.basex.org/wiki/Options#DEFAULTDB