Gives KeyError after deleting object from dictionary

[PlayfulPiano]

Type: Bug

Behaviour

Expected vs. Actual

Deleting Key/Value Pair would not affect DW viewing
Gives KeyError

Steps to reproduce:

1. Create Dictionary.
2. View object using DW in Python Debugger. No issues.
3. delete value in dictionary using del dict[key]
4. View object using DW in Python Debugger. Gives KeyError.

Diagnostic data

• Jupyter extension version: 2024.11.0
• Python extension version: 2025.0.0
• .NET Install Tool for Extension Authors extension version: Not installed
• Python package dependencies:

{
  "installed": {
    "pandas": "2.2.2",
    "pyarrow": "19.0.0"
  },
  "required": {
    "pandas": "1.2.0"
  },
  "unsatisfied": []
}

• Entrypoint: JupyterDataViewer
• Active mode: dataViewer

Extension version: 1.19.1
VS Code version: Code 1.97.0 (33fc5a94a3f99ebe7087e8fe79fbe1d37a251016, 2025-02-04T22:41:26.688Z)
OS version: Windows_NT x64 10.0.22631
Modes:

System Info

Item	Value
CPUs	13th Gen Intel(R) Core(TM) i5-1345U (12 x 2496)
GPU Status	2d_canvas: enabled canvas_oop_rasterization: enabled_on direct_rendering_display_compositor: disabled_off_ok gpu_compositing: enabled multiple_raster_threads: enabled_on opengl: enabled_on rasterization: enabled raw_draw: disabled_off_ok skia_graphite: disabled_off video_decode: enabled video_encode: enabled vulkan: disabled_off webgl: enabled webgl2: enabled webgpu: enabled webnn: disabled_off
Load (avg)	undefined
Memory (System)	31.64GB (7.92GB free)
Process Argv	--crash-reporter-id 0550736f-c139-4c88-9107-9ac488208fed
Screen Reader	no
VM	0%

A/B Experiments

vsliv368cf:30146710
vspor879:30202332
vspor708:30202333
vspor363:30204092
vswsl492cf:30256860
vscod805:30301674
binariesv615:30325510
vsaa593cf:30376535
py29gd2263:31024239
14231c:31235034
c4g48928:30535728
azure-dev_surveyone:30548225
2i9eh265:30646982
962ge761:30959799
h48ei257:31000450
pythontbext0:30879054
cppperfnew:31000557
dwnewjupyter:31046869
nativerepl1:31139838
pythonrstrctxt:31112756
nativeloc1:31192215
iacca1:31171482
5fd0e150:31155592
dwcopilot:31170013
stablechunks:31184530
6074i472:31201624
hdaa2157:31222309
copilot_t_ci:31222730
jda6j935:31233686

[PlayfulPiano]

Code used to delete keys:

for idx in list(standardizations_dict.keys()):
    if len(standardizations_dict[idx]) > 1:
        single_row_keys = [key for key in standardizations_dict[idx] if len(standardizations_dict[idx][key]) == 1]
        for key in single_row_keys:
            del standardizations_dict[idx][key]

[PlayfulPiano]

Notable update: KeyError does not occur if the dictionary object is viewed in Data Wrangler prior to the aforementioned code, followed by refreshing the data within Data Wrangler after said code concludes. So the KeyError only occurs if Data Wrangler did not view the object before deleting key/value pairs in the dictionary.

[pwang347]

Hi @PlayfulPiano, thank you for opening this issue!

I tried running the following code based on your description but did not run into the issue:

standardizations_dict = {
    1: {
        "a": ["a"],
        "b": ["b", "c"],
    },
    2: {
        "a": ["b"],
        "b": ["c", "d"],
    },
}

for idx in list(standardizations_dict.keys()):
    if len(standardizations_dict[idx]) > 1:
        single_row_keys = [key for key in standardizations_dict[idx] if len(standardizations_dict[idx][key]) == 1]
        for key in single_row_keys:
            del standardizations_dict[idx][key]

print(standardizations_dict) # breakpoint here and view `standardizations_dict`

standardizations_dict loads correctly here at the breakpoint although the key was deleted with the loop above:

Some questions/followups:

1. Am I correct in understanding that you are getting a KeyError by just opening at the breakpoint, and not after refreshing?
2. Would you be able to modify the code above to recreate the error?
3. Could you please share a screenshot of the KeyError?

Thanks!

[PlayfulPiano]

Hi @PlayfulPiano, thank you for opening this issue!

I tried running the following code based on your description but did not run into the issue:

standardizations_dict = {
1: {
"a": ["a"],
"b": ["b", "c"],
},
2: {
"a": ["b"],
"b": ["c", "d"],
},
}

for idx in list(standardizations_dict.keys()):
if len(standardizations_dict[idx]) > 1:
single_row_keys = [key for key in standardizations_dict[idx] if len(standardizations_dict[idx][key]) == 1]
for key in single_row_keys:
del standardizations_dict[idx][key]

print(standardizations_dict) # breakpoint here and view standardizations_dict

standardizations_dict loads correctly here at the breakpoint although the key was deleted with the loop above:

Some questions/followups:

1. Am I correct in understanding that you are getting a KeyError by just opening at the breakpoint, and not after refreshing?

2. Would you be able to modify the code above to recreate the error?

3. Could you please share a screenshot of the KeyError?

Thanks!

1. If I view the dictionary with Data Wrangler via breakpoint prior to the key deletion excerpt, then refresh it after the key deletion excerpt finishes via breakpoint, I will not receive a KeyError. If I have never viewed the dictionary with Data Wrangler prior to the key deletion excerpt, then attempt to view it after the excerpt finishes via breakpoint, then I will receive a KeyError.

As for 2/3, I can provide that tomorrow mid-day as I will have access to the code at that point.

[PlayfulPiano]

Code: https://gist.github.com/PlayfulPiano/3de000a396b46c9ce25dc5a4f1e62c17

[PlayfulPiano]

Just want to quickly mention that the newest prerelease (1.19.2) still has the same issue:

As well as in the release (1.18.2):

[pwang347]

Thank you for the extra details @PlayfulPiano - unfortunately I've tested this in several environments and have been unable to reproduce the same issue with the code you've shared above.

Just to double check, could you please share the following information with us? Much thanks in advance!

• Python interpreter version
• Debugpy python package version
• IPython python package version (if it is installed)
• Does the same issue occur for you while debugging in a .ipynb file using the same Python interpreter?
• Do you have any other Python environments you can try, and does it error at the same location?
• Do you have any custom configs for your Python debugger?
• Is this how you are starting the Python debugger (in a .py file?)

• If possible, could you please enable debugger logging with the instructions here, trigger the error, and then copy the debugpy.pydevd results here? (note: I would advise disabling this later so you don't create too much junk files on your disk)

Thanks!

[PlayfulPiano]

Thank you for the extra details @PlayfulPiano - unfortunately I've tested this in several environments and have been unable to reproduce the same issue with the code you've shared above.

Just to double check, could you please share the following information with us? Much thanks in advance!

* Python interpreter version

* Debugpy python package version

* IPython python package version (if it is installed)

* Does the same issue occur for you while debugging in a .ipynb file using the same Python interpreter?

* Do you have any other Python environments you can try, and does it error at the same location?

* Do you have any custom configs for your Python debugger?

* Is this how you are starting the Python debugger (in a .py file?)

* If possible, could you please enable debugger logging with the instructions [here](https://github.com/microsoft/debugpy/wiki/Enable-debugger-logs), trigger the error, and then copy the `debugpy.pydevd` results here? (note: I would advise disabling this later so you don't create too much junk files on your disk)

Thanks!

Hi sorry for the late response:

3.12.8

v4.5

N/A (I don't use jupyter)

I haven't tried that

I don't have other environments

Yes but mainly related to type hints

Yes.

[pwang347]

Hi @PlayfulPiano, no worries at all.

A few items that would help me investigate here:

• Could you please run import debugpy; print(debugpy.__version__) to get the version number? I'm expecting to see a number in this range: https://pypi.org/project/debugpy/#history
• If you don't mind, could you please share your debug config file?
• When you get the time, it would be great if you could try some of the items I mentioned (e.g., enabling debug logs, using a Jupyter notebook).

Thanks!

[PlayfulPiano]

debugpy version: 1.8.5
debug log: Data Wrangler Kernel Output (standardizations_dict).log
regarding jupyter, do you want me to just copy the entire script into a code cell?
as for the config file, are you referring to this?:

{
    "security.workspace.trust.untrustedFiles": "open",
    "files.associations": {
		"*.rmd": "markdown"
	},
    "debug.javascript.resourceRequestOptions": {},
    "win-ca.inject": "append",
    "github.copilot.advanced": {
		"authProvider": "github"
	},
    "workbench.editorAssociations": {
		"*.copilotmd": "vscode.markdown.preview.editor",
		"{hexdiff}:/**/*.*": "hexEditor.hexedit",
		"*.xlsm": "gc-excelviewer-excel-editor",
		"{git,gitlens,git-graph}:/**/*.{md,csv,svg}": "default",
		"*.austin": "speedscope-in-vscode.speedscope",
		"*.pdf": "cweijan.officeViewer",
		"*.md": "default"
	},
    "editor.guides.bracketPairs": "active",
    "editor.indentSize": "tabSize",
	"editor.tabSize": 4,
    "editor.detectIndentation": false,
	"editor.insertSpaces": false,
	"workbench.tree.renderIndentGuides": "always",
	"editor.bracketPairColorization.independentColorPoolPerBracketType": true,
	"editor.wordWrap": "on",
	"chat.editor.wordWrap": "on",
	"python.defaultInterpreterPath": "C:\\Users\\VGS13\\AppData\\Local\\Microsoft\\WindowsApps\\python3.12.exe",
	"vscode-office.editorTheme": "Github Dark",
	"vscode-office.openOutline": false,
	"files.autoGuessEncoding": true,
	"diffEditor.ignoreTrimWhitespace": true,
	"diffEditor.hideUnchangedRegions.enabled": true,
	"debug.toolBarLocation": "docked",
	"dataWrangler.enabledFileTypes": {
		"jsonl": true
	},
	"dataWrangler.useRelativePath": true,
	"github.copilot.chat.followUps": "always",
	"github.copilot.chat.inlineChatHint.enabled": true,
	"github.copilot.chat.setupTests.enabled": false,
	"python.analysis.autoFormatStrings": true,
	"python.analysis.autoImportCompletions": true,
	"python.analysis.inlayHints.callArgumentNames": "off",
	"python.analysis.inlayHints.functionReturnTypes": false,
	"python.analysis.inlayHints.variableTypes": false,
	"python.analysis.inlayHints.pytestParameters": false,
	"python.analysis.generateWithTypeAnnotation": false,
	"pythonIndent.trimLinesWithOnlyWhitespace": true,
	"python.createEnvironment.trigger": "off",
	"livecode.pythonPath": "C:\\Program Files\\WindowsApps\\PythonSoftwareFoundation.Python.3.12_3.12.2288.0_x64__qbz5n2kfra8p0\\python3.12.exe",
	"inlineChat.accessibleDiffView": "on",
	"inlineChat.mode": "preview",
	"austin.binaryMode": true,
	"austin.lineStats": "Both",
	"cSpell.languageSettings": [
		{
            "languageId": "python,cpp,c",
            "includeRegExpList": [
                // For Python
                "comments",
                "strings",
                // For C++ and C
                "CStyleComment",
                "string"
            ]
        },
        {
            "languageId": "jsonc", // (.json with comments) because most "commands" here are strings
            "includeRegExpList": [
                "CStyleComment",
            ]
        }
	],
	"cSpell.userWords": [
		"APPY",
		"Backflush",
		"dataframes",
		"isna",
		"matr",
		"mult",
		"pretasks",
		"SAPB",
		"subdf",
		"subpre"
	],
	"dataWrangler.experiments.fastCsvParsing": true,
	"dataWrangler.experiments.autoDetectCsvDelimiter": true,
	"errorLens.codeLensEnabled": true,
	"errorLens.scrollbarHackEnabled": true,
	"errorLens.codeLensTemplate": "$severity $message $source",
	"errorLens.messageEnabled": false,
	"errorLens.codeLensOnClick": "showProblemsView",
	"errorLens.decorations": {
		"errorRange": {},
		"warningRange": {},
		"infoRange": {},
		"hintRange": {},
		"errorMessage": {},
		"warningMessage": {},
		"infoMessage": {},
		"hintMessage": {}
	},
	"cSpell.patterns": [],
	"cSpell.diagnosticLevel": "Hint",
	"mypy-type-checker.importStrategy": "fromEnvironment",
	"mypy-type-checker.args": [
		"--allow-untyped-globals",
		"--disallow-incomplete-defs",
		"--check-untyped-defs"
	],
	"errorLens.lintFilePaths": {
		"eslint": [
			"**/*.eslintrc.{js,cjs,yaml,yml,json}",
			"**/*package.json"
		],
		"Stylelint": [
			"**/*.stylelintrc",
			"**/*.stylelintrc.{cjs,js,json,yaml,yml}",
			"**/*stylelint.config.{cjs,js}",
			"**/*package.json"
		]
	},
	"mypy-type-checker.preferDaemon": true,
	"mypy-type-checker.severity": {
		"error": "Warning"
	},
	"vscode-note.notespath": "Documents",
	"python.analysis.typeEvaluation.strictDictionaryInference": true,
	"python.analysis.typeEvaluation.strictListInference": true,
	"python.analysis.typeEvaluation.strictSetInference": true,
	"dataWrangler.defaultViewingFilterOptions": "advanced",
	"dataWrangler.outputRenderer.columnInsights.displayByDefault": true,
	"debug.inlineValues": "off"
}

[pwang347]

Hi @PlayfulPiano, thank you for providing the extra details, here are some clarifications below:

debug log: Data Wrangler Kernel Output (standardizations_dict).log

Apologies, I meant this one actually: https://github.com/microsoft/debugpy/wiki/Enable-debugger-logs (I personally use the environment variable option, since it's a bit easier to add/remove and you can customize which folder the logs go to).
And then sharing the contents of a file called debugpy.pydevd.<number>.log

regarding jupyter, do you want me to just copy the entire script into a code cell?

Yes, specifically it would be great if you could try the following:

1. Paste script into a code cell
2. Add breakpoint at the same location
3. Debug the cell

as for the config file, are you referring to this?:
Actually, I meant the .vscode/launch.json file described here: https://code.visualstudio.com/docs/python/debugging#_initialize-configurations

If you don't have any custom parameters being sent to the debugger it's probably not important. Thanks again!

[PlayfulPiano]

debugpy.pydevd.12796.log

datawrangler telemetry.txt

doesn't appear to cause issues within jupyter as far as I can tell, and I don't have any custom parameters before for the debugger

[pwang347]

Hi @PlayfulPiano! Thanks to the information you provided, our team found a bug in Python 3.12 :)

Just wanted to let you know that this issue should now be resolved on our side in the latest versions (1.20.1 and 1.21.1) of Data Wrangler and a fix in Python 3.12 itself is currently being discussed/pending.

It would be great if you could update and let us know if you're still running into this. Thank you again for the patience and for helping us with the investigation!