Add tfio.IOTensor.from_csv support (experimental with Apache Arrow's CSV parser) #443
Conversation
This file contains hidden or 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
yongtang
force-pushed
the
csv_io_tensor
branch
2 times, most recently
from
f4ea342 to
6cbca6c
Compare
yongtang
force-pushed
the
csv_io_tensor
branch
from
6cbca6c to
8f51f3a
Compare
…CSV parser) CSV file is one of the most widely used format and in TensorFlow's main repo there is already a CsvDataset which could be conviniently used for iteration, either feed into tf.keras directly, or access through a for loop. There are, still some reasons to have a CSV input processor that give indexing and slicing access. The most notable reason is that while CSV file itself technically is only splittable (not truly indexable), in reality especially in data science by default CSV file is almost always loaded into memory. And because of its wide usage, it is really more convenient and more flexible to have a CSV processor that allows indexing and slicing. This PR takes the indexing and slicing approach and built the parser on top of Arrow. One advantage of Arrow is that Arrow's CSV's parser options are closer to widely used pandas. This will allow easy usage of importing csv files created by pandas. Signed-off-by: Yong Tang <[email protected]>
Signed-off-by: Yong Tang <[email protected]>
yongtang
force-pushed
the
csv_io_tensor
branch
from
40967b5 to
4e9ec5f
Compare
|
Plan to merge this PR soon as well, it largely follows existing patterns. Could address in follow up PRs if any issue surface. |
i-ony
pushed a commit
to i-ony/io
that referenced
this pull request
Feb 8, 2021
…CSV parser) (tensorflow#443) * Add tfio.IOTensor.from_csv support (experimental with Apache Arrow's CSV parser) CSV file is one of the most widely used format and in TensorFlow's main repo there is already a CsvDataset which could be conviniently used for iteration, either feed into tf.keras directly, or access through a for loop. There are, still some reasons to have a CSV input processor that give indexing and slicing access. The most notable reason is that while CSV file itself technically is only splittable (not truly indexable), in reality especially in data science by default CSV file is almost always loaded into memory. And because of its wide usage, it is really more convenient and more flexible to have a CSV processor that allows indexing and slicing. This PR takes the indexing and slicing approach and built the parser on top of Arrow. One advantage of Arrow is that Arrow's CSV's parser options are closer to widely used pandas. This will allow easy usage of importing csv files created by pandas. Signed-off-by: Yong Tang <[email protected]> * Fix python 3 error Signed-off-by: Yong Tang <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
CSV file is one of the most widely used format and in TensorFlow's main repo there is already a CsvDataset which could be conviniently used for iteration, either feed into tf.keras directly, or access through a for loop.
There are, still some reasons to have a CSV input processor that give indexing and slicing access. The most notable reason is that while CSV file itself technically is only splittable (not truly indexable), in reality especially in data science by default CSV file is almost always loaded into memory. And because of its wide usage, it is really more convenient and more flexible to have a CSV processor that allows indexing and slicing.
This PR takes the indexing and slicing approach and built the parser on top of Arrow. One advantage of Arrow is that Arrow's CSV's parser options are closer to widely used pandas. This will allow easy usage of importing csv files created by pandas.
Signed-off-by: Yong Tang [email protected]