Refactor FormattedVector for faster formatting of S3 objects
#646
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
dfalbel
force-pushed
the
bugfix/refactor-formatted-vector
branch
from
15f6f34 to
897cc9f
Compare
dfalbel
force-pushed
the
bugfix/refactor-formatted-vector
branch
from
897cc9f to
19fe643
Compare
lionel-
approved these changes
Dec 9, 2024
| /// Returns an iterator over the first `n` elements of a vector. | ||
| /// Should be used when the vector is potentially large and you won't need to | ||
| /// iterate over the entire vector. | ||
| pub fn iter_n(&self, n: usize) -> anyhow::Result<FormattedVectorIter> { |
| /// If `indices` is `None`, the iterator will iterate over all elements of the vector. | ||
| /// If `indices` is `Some`, the iterator will iterate over the elements at the specified indices. | ||
| /// The indices must be valid and in bounds as we do no bounds checking. | ||
| fn new_unchecked( |
There was a problem hiding this comment.
This method is missing a Safety comment to explain what is being unchecked exactly and what are the failure modes (panics, R errors?).
There was a problem hiding this comment.
Added some notices here: e862651
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The variables pane uses the
FormattedVectorAPI to format vector elements to be displayed. One important assumption that we do when using this API is that it's lazy, in the sense that we create an iterator over the elements of a vector and we only pay the formatting price for the values we actually iterated.This is true for atomic vectors, for which we iterate using R's C API eg, extracting values with
REAL_ELTand formattinng them in Rust. However, this is not true for any vector that has a class (except factors), in this case we need to use theformatfunction, and we were actually formatting the entire vector before being able to iterate over the formatted values.This is very costly if we need for format very large vectors, as when we expand the data.frame in posit-dev/positron#3628 (comment)
This PR refactors the
FormattedVectorAPI, with the main change being:FormattedVectoriter()iter(),iter_n(),column_iter(),column_iter_n(). Those suffixed withnwill only format the requested part of the vector.As a consequence, now creating the
iter()returns aResult, so we need to adapt in a few places.Now, most of the times in the variables pane we'll use
iter_n()to only format the firstNelements, that are necessaryto create the display value in the variables pane.