Import table issues

[hansfn]

Info: Kirix 4.0 beta 4 on Win 2000 Pro.

1) The "Import Table" entry in the context/right click menu, is still present even if there are no tables present in the web page. This has been reported earlier, and might be (forgotten) on the feature request list?

2) It should be possible to import all tables in one go. At least, I wasn't able to do so - I tried select all and then import table.

3) All tables should be (optionally) marked when viewing a webpage or at least the one you have currently selected. I could have done this myself (if there was an easy way to add user style sheets), but I think this should be builtin.

4) When importing tables with numerical data that contains "." or ",", Kirix must ask what the decimal separator is. I just imported "7,50" (seven and a half) and got 750 ... Optionally the default decimal separator, could be added to the Options.

5) Whitespace handling: White space in HTML is collapsed. Kirix doesn't do this so a cell that looks like a normally spaced sentences in a browser, might become very "ugly" in Kirix. I suggest that also Kirix collapse whitespace inside each table cell.

Enough complaints: I was very happy to see that Kirix had no problems with important characters outside US-ASCII - for example the Norwegian letters ÆØÅ. Thx for a great product.

[Ken]

Hi there, thanks for the great feedback!

Some responses and a couple questions for clarification:

1) As noted, this is on our bug list. It has just been lower priority compared to some of the other areas we are wrestling down at the moment. I expect that we'll have this sorted before the "final" release.

2) At this point, some file types can be imported in one go (e.g., csv files) and others cannot (e.g., xls files). Also, you can't import "mixed" file types (one csv file and one xls file) together. What type of file(s) where you wanting to import together?

There is also some lack of clarity in the import interface that may have thrown you. Say you browsed your desktop and selected 3 csv files. When you hit "OK", the "Files" in the import dialog show only a single path. The differentiator, at this point, between one file and multiple files is that multiple files have quotation marks (") around it. But, if you hit next to select the delimiter and then next again, you'll see all three files ready to import.

[EDIT: After a re-read, you may have just meant "import all tables on a given web page" instead of importing from your desktop/server. If that's the case, I'll add it to the feature list. Do you have a particular web page in mind that we could see as an example? Thanks.]

3) By "marked," do you mean that, for instance, every importable table on a web page should have a, say, red box around it to help show the user what they can click on? This is something that we've discussed and would like to do at some point -- it particularly is useful for selecting only part of an HTML table. I don't think it is going to happen for this first version, but it is a feature we are considering. Please let me know if you are thinking about something else though... we're always happy to hear about usability improvement ideas.

4) Yes, this is a bug; we still have this formatting issue to work out. Also, glad to hear that your Unicode text was importing well. However, there are still a few more issues with unicode that we are still working out.

5) Regarding whitespace, could you provide an example of what you are referring to... I'm not quite sure I understand (it could be a bug). Whitespace in any given HTML markup should render like it would in FireFox or IE. In tables, you can control whitespace by selecting on the column/row handle and moving it (like you would in Excel). However, my hunch is that you may have imported a table that is not being parsed correctly which padded the cell with extra "space" characters. Not sure though -- if you can point me to a web page or send me a file that would be great. If it is private, you can send me a PM or an email to support[at]kirix.com.

Glad you are enjoying the product, despite some of the existing bugs. As you encounter more issues, please let us know -- these types of posts are extremely valuable. Thanks!

#61, #439, #446, #447

[hansfn]

1) I'm glad it's still on the list. I do understand that other issue have higher priorities.

2 and 3) Yes, I meant "import all tables on a given web page". And yes, by marked I meant for example a red border around the table. By the way, I don't think selecting a part of a HTML table is very useful - you can easily remove the unwanted columns/rows in the data part of Kirix.
Example: http://himolde.studiehandbok.no/no/cont ... /full/9943 (In Norwegian, sorry.)
This example is nice because if you add "table.emnetabell{ border: red dashed 2px;}" to the site style sheets (using the Web Developer or Firebug extension or User style sheets), you'll automatically a) see all the tables and b) see how I think a selected table can be marked (easily).

5) Try importing the very simple table at http://home.himolde.no/~nordhaug/programvare/ After importing you'll find in the "Beskrivelse" (description) column stuff like: "[...] jeg valgte fordi Rational Rose [...]". Notice the spaces between "valgte" and "fordi"? (I guess not because the web browser will collapse it - damn.) Those are collapsed to one space in the browser and should be one space in the data part of Kirix too. (If I had used multiple nbsp it would have been a different story.)

Thx for the quick reply.

[Ken]

2 and 3) Thanks for the link and the mark up... we'll definitely throw it on the feature list. I would love to have you just create a quick stylesheet override extension... but unfortunately we haven't implemented the functionality to get at the stylesheet just yet. It is something we'd like to do in the future though.

Regarding deleting rows/columns and where this "mark up" may come in handy: The annoying issue here is when you have a perfectly nice HTML table, but the creator of the table puts in multiple header rows. Something like this:

Meta 1, Meta 2, Meta 3
Date 1, Date 2
Field1, Field2, Field3, Field4, Field5, Field 6
abc, 123, 11/06/07, abc, 123, a1b2
etc

When Strata auto-senses the table, it can't always tell which is the true field header row and therefore then cannot interpret the field types correctly. So, in the case above, you may have to 1) delete an extraneous couple rows 2) rename the fields and 3) reset the types. It is completely doable manually -- but it would be nice to be able to automate the process to select only the part of the table you care about. It is a tricky thing to do though since the web is such a "wild" place, but we'll be working on it.

5) Thanks very much for the example; I see exactly what you mean. I'll throw this onto the bug list.

#458, #459, #447, #460