Question: save and export options?

[jools-r]

Once you've parsed, filtered, and concatenated various data sources, are there any ways to export the data results (and/or maps, diagrams, charts) or to build new databases from them? If so, what formats are supported?

A related question: is it possible to save data pipelines for re-using on different data sources of the same format?

Currently the homepage is unclear on both. A test version would make that apparent ;-)

[kengruven]

Thanks for writing! A webpage update is on my short list, so I'll be sure to add some clarification around these points you brought up.

• Right now, there are two basic ways to get data out of Strukt: writing back to a file (text or binary), or copying to the clipboard. Many of the file formats that can be parsed by Strukt can also be generated (like CSV, TSV, HTML, XML, and JSON). Some can't, yet (like XLSX). It's also possible to build-your-own other formats by hand, using operations. (Obviously, this is much more feasible for some formats than others.)

  Writeback to databases isn't supported, yet, either, but it's on my list. Most databases support CSV import, so you could always go that route.

• There's no way to export the maps/charts/etc directly. (You can always, of course, take a screenshot.) My thinking is that for publishing, you'll generally want to put the raw data into the target application (Excel/Word, Numbers/Keynote, ...) and make the charts there, so you can match the style of the rest of your document. But I can see how it could be useful to generate charts directly. What format(s) would you want?

• It is possible to save data pipelines to a file, and re-open and re-run them later. It supports the standard set of macOS document operations, so you can easily "Duplicate" a pipeline and replace one data source operation with another, for example. Or you can put two "load" operations next to each other in the same pipeline, and alternately disable just one.

For all of the things it doesn't support yet, the only reason is they haven't been the top priority, and/or nobody has ever thought to ask for them before. I do add features to Strukt all the time (I added 2 operations just last week!), so if there's a particular feature you need, just ask. It might be easy to implement, or something I'm working on already. I can totally appreciate asking somewhat vague questions, to feel out what an app can do, but the functionality is improving all the time, and if you're looking for a specific feature, there's a good chance someone else is, too!

cheers!

[jools-r]

Thank you for the very informative answer. That was exactly what I wanted to know about both aspects.

CSV is fine as a bridge to other common DB formats so long as the individual fields remain distinct. I was envisaging using Strukt to extract legacy data in many small files for use in a new CMS, and being able to export is vital for that workflow, as is reusing data pipelines once I have them working. There are probably some niggles to deal with such as escaping certain characters for CSV format or CDATA blocks. Sometimes it's useful to perform string operations on 'converted' output (make lowercase or url-friendly, etc., prepend/append a string), but some of that can be done in excel/numbers.

Regarding chart/diagrams, it can be useful to have diagrams in a vector format that one can post-edit, clean up, adapt etc. Even without a specific vector export format (e.g. EPS), one can often dissect a PDF printout if the content it contains is in vector-based objects. Personally, I would not see this a priority.

[kengruven]

I was envisaging using Strukt to extract legacy data in many small files for use in a new CMS

I don't think I've heard of anyone using it for that, but it certainly sounds like something that I want Strukt to be able to do. If you run into any limitations, let me know and I'll be happy to help.

Sometimes it's useful to perform string operations on 'converted' output (make lowercase or url-friendly, etc., prepend/append a string), but some of that can be done in excel/numbers.

FWIW, all of these can be done in Strukt:

one can often dissect a PDF printout if the content it contains is in vector-based objects. Personally, I would not see this a priority.

Good points. I'll keep this in mind. The maps use Apple's MapKit, which doesn't seem to support any vector output, but it could be done for the charts. One possible method could be to add printing support for output views -- then you could just save to PDF instead.

[molschimke]

+1 for writing data back into database (in our case especially PostgreSQL and MongoDB)