feat: Add `DataFrame.from_dict` by dangotbanned · Pull Request #2885 · narwhals-dev/narwhals

dangotbanned · 2025-07-25T18:19:01Z

What type of PR is this? (check all applicable)

Related issues

Part of API: io functions for v2 #2116

Checklist

Code follows style guide (ruff)
Tests added
Documented the changes

If you have comments or can explain your changes, please do so below

Resolve [var-annotated] by not using Self (695f9ac)

Possible Follow-ups

Figure out how (Establish safe patterns using Implementation.UNKNOWN #2786) fits in
Add @overloads on backend
- Like in Namespace
Defer to DataFrame.from_dict in nw.from_dict

Towards #2116

Will fix after running ci

dangotbanned · 2025-07-25T18:30:57Z

narwhals/dataframe.py

            ```
    """

+    _version: ClassVar[Version] = Version.MAIN


@MarcoGorelli

Doing this is gonna be quite handy for the rest of (#2116)

Since we're already inside the class, we can just start from the known version there

narwhals/narwhals/dataframe.py

Line 512 in 695f9ac

ns = cls._version.namespace.from_backend(implementation).compliant

instead of using _stablify afterwards 😄

narwhals/narwhals/functions.py

Line 316 in f8f7f65

ns = Version.MAIN.namespace.from_backend(implementation).compliant

#2885 (comment)

- Part of #2116 - Based on #2885

MarcoGorelli

thanks @dangotbanned !

just left a comment

MarcoGorelli · 2025-07-26T12:16:13Z

narwhals/dataframe.py

+        if is_eager_allowed(implementation):
+            ns = cls._version.namespace.from_backend(implementation).compliant
+            compliant = ns._dataframe.from_dict(data, schema=schema, context=ns)
+            return cls(compliant, level="full")
+        # NOTE: (#2786) needs resolving for extensions
+        msg = (
+            f"{implementation} support in Narwhals is lazy-only, but `DataFrame.from_dict` is an eager-only function.\n\n"
+            "Hint: you may want to use an eager backend and then call `.lazy`, e.g.:\n\n"
+            f"    nw.DataFrame.from_dict({{'a': [1, 2]}}, backend='pyarrow').lazy('{implementation}')"
+        )
+        raise ValueError(msg)


i'm confused as to why is this necessary - if we have a DataFrame class, aren't we already eager?

I was just adapting the existing error in nw.from_dict

narwhals/narwhals/functions.py

Lines 328 to 332 in ee06d33

msg = (

f"{implementation} support in Narwhals is lazy-only, but `from_dict` is an eager-only function.\n\n"

"Hint: you may want to use an eager backend and then call `.lazy`, e.g.:\n\n"

f" nw.from_dict({{'a': [1, 2]}}, backend='pyarrow').lazy('{implementation}')"

)

Did you want this version to do something different if we wrote?

import narwhals as nw nw.DataFrame.from_dict({'a': [1, 2]}, backend="pyspark")

ah I see sorry, this is indeed needed and correct, thanks!

my only other comment would be - do we want to repeat the docstring, or do like numpy and say "see [...] for full docstring"? probably fine to repeat?

https://numpy.org/doc/stable/reference/generated/numpy.ndarray.std.html

"see [...] for full docstring"? probably fine to repeat?

Definitely repeat!

I'm guessing even to get that doc you'd need to not have type stubs or have a jupyter kernel running?

I'm always gonna feel strongly about IDE-experience-first-docs, since that's how I consume them 🙂

Sadly useful docs + static typing often mean we need to repeat ourselves.
But putting in that extra effort is something I really appreciate from polars vs pandas

MarcoGorelli

this needs adding to dataframe.md in the docs, right?

other than that, looks good

dangotbanned · 2025-07-26T13:14:40Z

this needs adding to dataframe.md in the docs, right?

oop yeah will do that now!

Resolves #2885 (review)

dangotbanned · 2025-07-26T13:56:53Z

(#2885 (comment))
Just need to fix check_api_reference.py to handle classmethod

Preview of the doc

Resolves https://results.pre-commit.ci/run/github/760058710/1753537609.xO2JX1vOTs2Wfs5FLE-1-w

dangotbanned added 4 commits July 25, 2025 17:20

feat: Add DataFrame.from_dict

2b40ede

Towards #2116

docs: Adapt from_dict docstring

5562b00

test: Adapt from_dict tests

24997d2

chore(typing): Ignore var-annotated for now

b523edc

Will fix after running ci

dangotbanned added the enhancement New feature or request label Jul 25, 2025

fix(typing): Resolve [var-annotated] by not using Self

695f9ac

dangotbanned commented Jul 25, 2025

View reviewed changes

test: Cover v1

bbefe38

#2885 (comment)

dangotbanned marked this pull request as ready for review July 25, 2025 18:42

dangotbanned mentioned this pull request Jul 25, 2025

API: io functions for v2 #2116

Open

6 tasks

dangotbanned added a commit that referenced this pull request Jul 25, 2025

feat: Add DataFrame.from_numpy

92aa433

- Part of #2116 - Based on #2885

This was referenced Jul 25, 2025

feat: Add DataFrame.from_numpy #2887

Merged

feat: Add DataFrame.from_arrow #2891

Merged

MarcoGorelli reviewed Jul 26, 2025

View reviewed changes

MarcoGorelli approved these changes Jul 26, 2025

View reviewed changes

dangotbanned added 2 commits July 26, 2025 13:43

Merge remote-tracking branch 'upstream/main' into dataframe-from-dict

ba1ea35

docs: add to api ref

55ba43a

Resolves #2885 (review)

dangotbanned marked this pull request as draft July 26, 2025 13:55

fix: Handle classmethod in check_api_reference.py

325f466

Resolves https://results.pre-commit.ci/run/github/760058710/1753537609.xO2JX1vOTs2Wfs5FLE-1-w

dangotbanned marked this pull request as ready for review July 26, 2025 14:07

dangotbanned merged commit bdcf661 into main Jul 26, 2025
33 checks passed

dangotbanned deleted the dataframe-from-dict branch July 26, 2025 15:35

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: Add `DataFrame.from_dict`#2885

feat: Add `DataFrame.from_dict`#2885
dangotbanned merged 9 commits intomainfrom
dataframe-from-dict

dangotbanned commented Jul 25, 2025 •

edited

Loading

Uh oh!

dangotbanned Jul 25, 2025

Uh oh!

MarcoGorelli Jul 25, 2025

Uh oh!

MarcoGorelli left a comment

Uh oh!

MarcoGorelli Jul 26, 2025

Uh oh!

dangotbanned Jul 26, 2025

Uh oh!

MarcoGorelli Jul 26, 2025

Uh oh!

dangotbanned Jul 26, 2025

Uh oh!

MarcoGorelli left a comment

Uh oh!

dangotbanned commented Jul 26, 2025

Uh oh!

dangotbanned commented Jul 26, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

	msg = (
	f"{implementation} support in Narwhals is lazy-only, but `from_dict` is an eager-only function.\n\n"
	"Hint: you may want to use an eager backend and then call `.lazy`, e.g.:\n\n"
	f" nw.from_dict({{'a': [1, 2]}}, backend='pyarrow').lazy('{implementation}')"
	)

Conversation

dangotbanned commented Jul 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What type of PR is this? (check all applicable)

Related issues

Checklist

If you have comments or can explain your changes, please do so below

Possible Follow-ups

Uh oh!

dangotbanned Jul 25, 2025

Choose a reason for hiding this comment

Uh oh!

MarcoGorelli Jul 25, 2025

Choose a reason for hiding this comment

Uh oh!

MarcoGorelli left a comment

Choose a reason for hiding this comment

Uh oh!

MarcoGorelli Jul 26, 2025

Choose a reason for hiding this comment

Uh oh!

dangotbanned Jul 26, 2025

Choose a reason for hiding this comment

Uh oh!

MarcoGorelli Jul 26, 2025

Choose a reason for hiding this comment

Uh oh!

dangotbanned Jul 26, 2025

Choose a reason for hiding this comment

Uh oh!

MarcoGorelli left a comment

Choose a reason for hiding this comment

Uh oh!

dangotbanned commented Jul 26, 2025

Uh oh!

dangotbanned commented Jul 26, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

dangotbanned commented Jul 25, 2025 •

edited

Loading