[New Plugin] CSV Lint

[Bas de Reuver]

@Thomas-McArthur and @Lycan-Thrope about the issue of CSV Lint and Notepad++ sometimes crashing when switching between different CSV tabs.

I think this is the same as issue #32, and recently I was able to reproduce a similar situation. That situation was due to potentially using an out-of-sync reference to the ScintillaGateWay and I’ve fixed it in this commit

Can you verify that it doesn’t crash anymore by using the latest development build? The syntax highlighting colors sometimes still don’t refresh immediately, only when you focus the file (click mouse somewhere in csv text), but at least it shouldn’t crash Notepad++ anywmore.

[Mark Olson]

@Bas-de-Reuver
I used to have problems with multi-instance/multi-view in JsonTools (I use an editor singleton global variable to interface with Scintilla), and what worked for me was to make it so that whenever the plugin receives an NPPN_BUFFERACTIVATED notification, the global editor is reassigned with editor = new ScintillaGateway(PluginBase.GetCurrentScintilla());.

In principle this approach should never fail, because that notification will always fire whenever the user toggles between instances/views. If one of the gurus can think of a reason why I’m wrong, I’d love to hear it.

Granted, you’d have to refactor all your existing code to remove all those calls, but it might simplify matters for you.

[Thomas McArthur]

@Bas-de-Reuver I tried the 64bit development build, but I get an error that it’s not compatible with v8.5.2. Are there dev builds of NP++? I don’t see any in the downloads page.

However, using your dev build may not be unnecessary because I can reproduce the problem in the non-dev 64bit version. The crash only happens when I have a CSV file that is too wide to fit on the page, so I turn off word wrap. With word wrap enabled, I can switch between CSV and non-CSV tabs fine. But when I enable word wrap in my 380-character wide CSV file, NP++ crashes when switching to that tab (and colorization is enabled).

[Mark Olson]

@Thomas-McArthur
The only way I know of to get a dev build of Notepad++ is to build it from source using Visual Studio.

I’ve tried CSVLint on my own 8.5.2 dev build and I can toggle word wrap on and off even with 400+ character columns on a 20MB file with no problems.

In my experience, it’s pretty rare for a clean build of Notepad++ to outright crash, so consider checking if any of your other plugins could be causing this.

[Thomas McArthur]

@Bas-de-Reuver v0.4.6.3 still crashes NP++.

I also tried downgrading NP++ to v8.4.5 and v8.5, and CSV Lint v0.4.6.3 still crashes those versions.

[Bas de Reuver]

@Thomas-McArthur That does sound like an serious issue, however I can’t reproduce what you’re describing.

I generated a test file with many columns (using generate_data99.py script) and then turning Word wrap on and off while switching between csv/non-csv tabs.

Maybe it has to do with the file encoding, does it also still happen when you first convert the file to a different encoding like ANSI or UTF8 and then save, and then try again? Or possibly it has to do with the session.xml file, does it still happen when you delete/rename your recent %userprofile%\AppData\Roaming\Notepad++\session.xml file and then re-open Notepad++? If it still happens, could you send me an e-mail with your ? > Debug info content?

[Bas de Reuver]

The CSV Lint plug-in was updated to v0.4.6.4 yesterday. The biggest new features are:

• Syntax highlighting improvement for Windows set to non-western codepage (Chinese, Korean etc.)
• Skip Lines / Comment Character for comment lines
• Dialogs have Dark mode support (preliminary)
• Improved support for quoted string
• Allow user to delete column
• Split on character, new option split on Nth occurrence
• Various bug fixes

The skiplines and comment lines feature:

And improved support for syntax highlighting when set to non-western codepage

I made the release, despite some reporting the new version crashing, because I wasn’t able to reproduce the issue (also see here and here).

If anyone can reproduce the plugin crashing on the latest Notepad++ version, or a possible cause, please let me know.

[Thomas McArthur]

@Bas-de-Reuver NP++ v8.5.2 64bit with CSV Lint v0.4.6.3 installed does not show any plugin updates.

[PeterJones]

@Thomas-McArthur said in [New Plugin] CSV Lint:

NP++ v8.5.2 64bit with CSV Lint v0.4.6.3 installed does not show any plugin updates.

Of course it doesn’t. Notepad++'s plugin list is only updated on a new release of Notepad++, and Notepad++ v8.5.2 came out on April 6, essentially a month before CSV Lint’s 0.4.6.4 update on May 2. If a plugin is released between Notepad++ versions and you want that updated plugin, you have to manually update the plugin.

[Thomas McArthur]

@Bas-de-Reuver I found the cause of the crash: it’s the Multi-Line setting for the Tab Bar. When that is enabled, and the tabs wrap to 2+ lines, activating a CSV tab a 2nd time crashes NP++.

• Enable multi-line, and only have one line of tabs - no crash.
• Un-maximize the window, drag it small enough to create 2+ rows of tabs - activating a CSV tab a 2nd time crashes NP++.

This crash even happens on a fresh download of NP++ v8.5.2 portable.

[rdipardo]

@Thomas-McArthur, I can’t make it crash for me, but it could simply be I don’t understand your instructions.

1. Does “Multi-Line” mean that long lines are wrapped, i.e. the option at View > Word wrap has a ✅ beside it?

2. Exactly how many different files are open at the same time? Is there a CSV file, and one more file of a different type? Or is the same CSV file open in multiple views, i.e., did you go to View > Move/Clone Current Document and then click Move/Clone to Other View?

[Snabel42]

@rdipardo I think this is checked: Settings / General / Tab Bar / Multi-line

Then make sure the window is narrow enough where the Tab bar has more than one line.

edit I can not replicate in version 0.4.6.4 64-bit

[Chr Pqe]

This plugin is a gamechanger. There are few bugs here and there nevertheless. For example,

• it will not properly detect documents in the second NPP pane.
• it seems to crash NPP a lot with CSV files using a lot of white or blank lines (maybe 50-100+ white/blank/tab lines at random positions)

I understand that this not a CSV or spreadsheet editor or a data analysis plugin, but some obvious features are missing :

• Change column position - add a simple menu item or button (in the CSV Lint window) to swap columns of the CSV
• Sort by this column - this is an obvious feature that is missing, nevertheless for this one, if the first feature is implemented (change column position), I could simply swap the needed column to the first one then use NPP sort and it should do the same thing.

I don’t think “changing column position” should be hard to integrate, since every column is already detected by a color (thus position).

Also, the reason it is a bit troublesome to not have this feature, is that a lot of times CSV columns are not in the correct position to visualize and understand the data, which forces to open Excel or some spreadsheet just to modify the position of the column and go back to NPP.

[Chr Pqe]

@artyb55 In your case for example, having a “change column position” would be useful aka you insert your column at the begening, then swap/change position of this newly edited column at the place you want

[Coises]

@Chr-Pqe said in [New Plugin] CSV Lint:

I understand that this not a CSV or spreadsheet editor or a data analysis plugin, but some obvious features are missing :

• Change column position - add a simple menu item or button (in the CSV Lint window) to swap columns of the CSV
• Sort by this column - this is an obvious feature that is missing, nevertheless for this one, if the first feature is implemented (change column position), I could simply swap the needed column to the first one then use NPP sort and it should do the same thing.

It is possible that a plugin I’m currently developing, Columns++, would help you, though there are some significant caveats:

1. It’s in alpha stage right now. I can’t promise it won’t crash Notepad++, so it’s important to keep work safe.
2. You would need to use its conversion functions to convert separated values to tabbed display, then convert back if you need to save the document in its original form. Depending on the details of your file, some care with the conversion options may be necessary if you need it to “round trip” unchanged except where you’ve made edits.
3. You would need to enable its Elastic tabstops feature to line up the columns; then you could use ordinary rectangular selections to move columns by cutting and pasting (admittedly, still a somewhat clumsy operation), and there are sort commands in the plugin that work with tabbed documents. Unfortunately, performance of Elastic tabstops is quite dependent on file size; e.g., I can edit a file with 1300 lines and 12 tabs per line and rarely notice any impact, but with 19,000 lines and 50 tabs per line, some operations can freeze for half a minute or more.

There’s a thread over here about it, so this one doesn’t get cluttered, if you have questions, comments or reports.

[Bas de Reuver]

@Chr-Pqe Thanks, nice to hear you like the plugin. 👍 Are there any functions of the plug-in that are especially helpful? I’d like to know because I can’t see which functions are most used or which aren’t used at all.

A sort function was in fact already added a while ago in v0.4.6, also see comment of 9 november 2022 in this thread, which version of the plug-in do you have installed? When you open the CSV Lint docking window Plugins > CSV Lint > CSV Lint window there should be a Sort button. It even sorts date/float/integer columns correctly according to their values.

And I agree, being able to change the column order would be a nice feature to have.

[Bas de Reuver]

@Chr-Pqe You say the plugin sometimes crashes, does that happen just when opening a csv file or when you do Reformat or something like that?

At the moment there is a known bug with the transparant cursor line, see issue #67 thanks to @Thomas-McArthur for posting it.

Does it still crash when you set the CSV Lint setting TransparantCursor to false?

[Bas de Reuver]

@Alan-Kilborn @rdipardo If possible, could you take a look at this issue #67 ? I’m 100% certain that this has to do with these two lines i.e. when I disabled them then it doesn’t crash anymore. However at the moment I can’t figure out why that is.

Btw the lines were introduced in this commit and also see this post

[Alan Kilborn]

@Bas-de-Reuver

It appears that @rdipardo had some thoughts on it, HERE.

[Bas de Reuver]

The CSV Lint plug-in was updated to v0.4.6.5. The biggest new feature is enumeration (coded values) meaning the plug-in now also tries to detect column with coded values. The big advantage is that you can now also check if a coded column contains invalid values, so values outside of the provided codelist. Other than that it’s mostly minor updates and bugfixes

• Support for enum/coded values
• Analyse Data Report, show correct date range
• Fix rare crashing situation due to transparent cursor line setting
• Syntax highlighting fix for special separator character
• Convert to SQL, PostgreSQL syntax improvements, support enum and table comment
• Generate Python/R script, support enum
• Generate W3C CSV Schema JSON various fixes and improvements
• Settings dialog, hide userpref dialog values
• Various updates and bugfixes