+1. The Thrift structures are private to parquet-cpp, so it should not
be an issue
On Wed, Sep 26, 2018 at 12:23 PM Zoltan Ivanfi
wrote:
>
> Hi,
>
> I just had a conversation with Nandor and he pointed out to me that even if
> we broke _reading_ in parquet-cpp 1.5.0, we could simply release a 1.5
Hi,
I just had a conversation with Nandor and he pointed out to me that even if
we broke _reading_ in parquet-cpp 1.5.0, we could simply release a 1.5.1
version that fixes it. The important thing is that _writing_ is good and
parquet-format-compliant in parquet-cpp 1.5.0, therefore we do not have
[
https://issues.apache.org/jira/browse/PARQUET-1425?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16628976#comment-16628976
]
Wes McKinney commented on PARQUET-1425:
---
I see, it's not a big deal. Perhaps ther
[
https://issues.apache.org/jira/browse/PARQUET-1425?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16628950#comment-16628950
]
Nandor Kollar commented on PARQUET-1425:
[~wesmckinn] would you like this to be
Hi,
Please let me know your opinions as well. So far all concerns were only
raised by me, which may indicate that other community members do not
consider this issue serious and in this case my suggestions may be
excessive and unjustified.
Just to clarify: The data structures for the encrypion fea
[
https://issues.apache.org/jira/browse/PARQUET-1427?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ASF GitHub Bot updated PARQUET-1427:
Labels: pull-request-available (was: )
> [C++] Move example executables and CLI tools to
Wes McKinney created PARQUET-1427:
-
Summary: [C++] Move example executables and CLI tools to Apache
Arrow repo
Key: PARQUET-1427
URL: https://issues.apache.org/jira/browse/PARQUET-1427
Project: Parque
Wes McKinney created PARQUET-1426:
-
Summary: [C++] parquet-dump-schema has poor usability
Key: PARQUET-1426
URL: https://issues.apache.org/jira/browse/PARQUET-1426
Project: Parquet
Issue Type
Wes McKinney created PARQUET-1425:
-
Summary: [Format] Fix Thrift compiler warning
Key: PARQUET-1425
URL: https://issues.apache.org/jira/browse/PARQUET-1425
Project: Parquet
Issue Type: Bug
[
https://issues.apache.org/jira/browse/PARQUET-1387?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Nandor Kollar updated PARQUET-1387:
---
Fix Version/s: format-2.6.0
> Nanosecond precision time and timestamp - parquet-format
> --
[
https://issues.apache.org/jira/browse/PARQUET-1266?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Nandor Kollar updated PARQUET-1266:
---
Fix Version/s: format-2.6.0
> LogicalTypes union in parquet-format doesn't include UUID
> -
[
https://issues.apache.org/jira/browse/PARQUET-1290?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Nandor Kollar updated PARQUET-1290:
---
Fix Version/s: format-2.6.0
> Clarify maximum run lengths for RLE encoding
> --
Oh, I see now. Moving the crypto structure from position 8 to 9. Let me
process that, and we can switch to a direct channel to continue this
discussion.
On Wed, Sep 26, 2018 at 3:30 PM Zoltan Ivanfi
wrote:
> Hi,
>
> I think it's safer if we skip id 8 altogether and use id 9 for the new
> crypto
Hi,
I think it's safer if we skip id 8 altogether and use id 9 for the new
crypto structure. This way we don't have to worry about remaining backwards
compatible with the accidentally released structure.
Br,
Zoltan
On Wed, Sep 26, 2018 at 2:25 PM Gidon Gershinsky wrote:
> Hi,
>
> I think we s
Hi,
I think we should use this id for its current purpose. This field had been
defined and merged months ago, and should be there is any scenario.
Except for the last week's change, the encryption format had been stable
for a while now. The timing of this change was unfortunate; but the change
was
Hi,
It seems that I spoke too early. I just noticed that a new field was added
to the ColumnChunk struct in
https://github.com/apache/parquet-cpp/pull/463/files#diff-0589b447b73e51c88d9b2bdcb0957084R708
Although the field is optional, we can't pretend it was never released,
because parquet-cpp alr
Sounds good. I agree a feature branch is the right thing to do in
these cases.
Cheers, Gidon.
On Wed, Sep 26, 2018 at 2:42 PM Zoltan Ivanfi
wrote:
> Hi,
>
> If the encryption code release in parquet-cpp is unused at this moment then
> I think we are fine. It means that we are still free to deci
Hi,
If the encryption code release in parquet-cpp is unused at this moment then
I think we are fine. It means that we are still free to decide any way
about the data structures without the risk of incompatility issues. In this
case I would suggest to proceed as we planned at the Parquet sync.
Tha
Hi Zoltan
PR #62 contains some rebase info which is not relate to change itself so I
created PR#99. Actually it only contains one file change now, I will add
another document file later.
Zoltan Ivanfi 于2018年9月26日周三 下午3:19写道:
> Hi,
>
> It seems to me that PR #99 does not supersede PR #62, as the
Hi,
It seems to me that PR #99 does not supersede PR #62, as the latter affects
16 files but the former only modifies a single one. Or has the rest of the
changes been already merged to the codebase from another PR? I checked the
history and I don't see anything related.
Thanks,
Zoltan
On Wed,
20 matches
Mail list logo