Skip to content

feat(parquet): Improved back-end storage for PageIndex - #11475

Merged
alamb merged 12 commits into
apache:mainfrom
etseidl:page_index_cache
Nov 13, 2026
Merged

alamb merged 12 commits into
apache:mainfrom
etseidl:page_index_cache

Conversation

@etseidl

@etseidl etseidl commented Oct 9, 2026 •

Copy link
Copy Markdown
Contributor

Which issue does this PR close?

Rationale for this change

The current PageIndex uses a nested vector of vectors to hold the page indexes. This wastes a great deal of memory when only a few columns or row groups are desired.

What changes are included in this PR?

  • Store populated indexes in compact, sorted slices keyed by packed row-group/column coordinates.
  • Parse indexes directly into row-major vectors, avoiding hash-map insertion and sorting in the default reader path.
  • Retain PageIndexBuilder for arbitrary insertion and replacement.
  • Add an example demonstrating an incremental page-index cache.

Are these changes tested?

Yes there are new tests added, but existing tests provide coverage as well.

Are there any significant user-facing changes?

There are no changes to the public API, but memory use for sparse indexes should drop.

Deprecates PageIndexBuilder::put_column_index and PageIndexBuilder::put_offset_index which silently fail in certain instances.

AI disclosure

Implementation, tests, and local benchmarks used coding-agent assistance, but I understand and own all the changes made.

@github-actions github-actions Bot added parquet Changes to the parquet crate enhancement New feature/addition labels Oct 9, 2026
@etseidl

etseidl commented Oct 9, 2026

Copy link
Copy Markdown
Contributor Author

cc @alamb and @adriangb when/if you have a moment.

@etseidl

etseidl commented Oct 9, 2026

Copy link
Copy Markdown
Contributor Author

It's also worth pointing out the 32-bit concerns that Codex noticed and documented at the head of page_index.rs.

assert_eq!(metadata.memory_size(), 8310);
#[cfg(feature = "encryption")]
assert_eq!(metadata.memory_size(), 11750);
assert_eq!(metadata.memory_size(), 8734);

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The added Arcs lead to some bloat for a fully populated index, but down here for a small subset the reduced storage overhead wins big.

@etseidl

etseidl commented Oct 9, 2026

Copy link
Copy Markdown
Contributor Author

run benchmark metadata

@adriangbot

Copy link
Copy Markdown

🤖 Benchmark starting (GKE) | trigger

Target: metadata

Sharding factor: 1 (1 workers).

Comparing page_index_cache (58ae317) to e189ba8 (merge-base) diff

Run configuration
run benchmark metadata
shards: 1

Results will be posted when all workers finish.


File an issue against this benchmark runner

@adriangbot

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Target: metadata

Comparing page_index_cache (58ae317) to e189ba8 (merge-base) diff

Run configuration
run benchmark metadata
shards: 1
Details
group                                               base                                   changed
-----                                               ----                                   -------
decode metadata (wide) with schema                  1.07     27.1±0.83ms        ? ?/sec    1.00     25.3±0.27ms        ? ?/sec
decode metadata (wide) with skip PES                1.08     27.7±0.88ms        ? ?/sec    1.00     25.6±0.25ms        ? ?/sec
decode metadata (wide) with skip all stats          1.03     30.9±0.75ms        ? ?/sec    1.00     30.1±0.22ms        ? ?/sec
decode metadata (wide) with skip column stats       1.06     28.7±0.87ms        ? ?/sec    1.00     27.1±0.22ms        ? ?/sec
decode metadata (wide) with skip size stats         1.04     31.6±0.82ms        ? ?/sec    1.00     30.4±0.26ms        ? ?/sec
decode metadata (wide) with stats mask              1.06     26.5±0.81ms        ? ?/sec    1.00     25.0±0.17ms        ? ?/sec
decode metadata with schema                         1.01      4.1±0.01µs        ? ?/sec    1.00      4.1±0.01µs        ? ?/sec
decode metadata with skip PES                       1.00      7.1±0.08µs        ? ?/sec    1.00      7.1±0.09µs        ? ?/sec
decode metadata with skip column stats              1.02      6.9±0.13µs        ? ?/sec    1.00      6.8±0.02µs        ? ?/sec
decode metadata with stats mask                     1.01      6.9±0.08µs        ? ?/sec    1.00      6.9±0.01µs        ? ?/sec
decode parquet metadata                             1.01      7.2±0.05µs        ? ?/sec    1.00      7.2±0.02µs        ? ?/sec
decode parquet metadata (wide)                      1.07     29.3±0.95ms        ? ?/sec    1.00     27.4±0.27ms        ? ?/sec
decode parquet metadata no path_in_schema (wide)    1.07     28.7±0.99ms        ? ?/sec    1.00     26.8±0.24ms        ? ?/sec
decode parquet metadata w/ size stats (wide)        1.06     36.0±0.79ms        ? ?/sec    1.00     33.8±0.19ms        ? ?/sec
open(default)                                       1.02      7.8±0.08µs        ? ?/sec    1.00      7.7±0.01µs        ? ?/sec
open(page index)                                    1.00    130.1±0.16µs        ? ?/sec    1.02    132.7±0.10µs        ? ?/sec
read page index                                     1.00    121.8±0.10µs        ? ?/sec    1.02    124.4±0.10µs        ? ?/sec
read page index reduced columns                     1.00     13.1±0.01µs        ? ?/sec    1.02     13.4±0.01µs        ? ?/sec

Per-runner information
metadata — shard 1/1

Node: gk3-benchmark-cluster-nap-1s1srbox-e6d80935-rrt4

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

uname:

Linux bench-c6087209513-3252-f4pq9 6.12.94+ #1 SMP Fri Aug 21 08:00:16 UTC 2026 aarch64 GNU/Linux

BENCH_COMMAND:

cargo bench --features=arrow,async,test_common,experimental,object_store --bench metadata
CPU Details (lscpu)
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Resource Usage

base (merge-base)

Metric Value
Wall time 175.0s
Peak memory 121.6 MiB
Avg memory 45.3 MiB
CPU user 171.9s
CPU sys 0.1s
Peak spill 0 B

branch

Metric Value
Wall time 170.0s
Peak memory 125.6 MiB
Avg memory 47.6 MiB
CPU user 168.1s
CPU sys 0.1s
Peak spill 0 B

File an issue against this benchmark runner

@adriangb

adriangb commented Oct 9, 2026

Copy link
Copy Markdown
Contributor

performance looks good!

@etseidl

etseidl commented Oct 9, 2026

Copy link
Copy Markdown
Contributor Author

performance looks good!

yeah, I'm mildly surprised the page index parsing didn't slow down more 😅

// specific language governing permissions and limitations
// under the License.

//! Sparse storage keyed by row-group and column coordinates.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

My current thinking is this can be reused for the Bloom filters, and perhaps ColumnChunkMetaData at some point.

Comment on lines +578 to +583
/// # Panics
///
/// Panics if either dimension exceeds `i32::MAX`, the maximum collection size representable
/// by Thrift.
///
/// Use [`try_new`](Self::try_new) for a fallible alternative.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is the biggest behavior change (along with the changes to the infallible allocate_ functions below). This used to take any valid usize dimension, but now panics if a dimension exceeds what's possible for Parquet.

That said, it may have panicked before anyway if it couldn't allocate a huge matrix.

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

allocating more than i32::MAX entries does sound like a bad idea to me and not a case that is important

@etseidl etseidl changed the title [WIP] Alternative back-end storage for PageIndex and PageIndexBuilder feat(parquet): Sparse back-end storage for PageIndex Oct 31, 2026
@etseidl etseidl changed the title feat(parquet): Sparse back-end storage for PageIndex feat(parquet): Improved back-end storage for PageIndex Oct 31, 2026
@etseidl

etseidl commented Oct 31, 2026

Copy link
Copy Markdown
Contributor Author

run benchmark metadata

@adriangbot

Copy link
Copy Markdown

🤖 Benchmark starting (GKE) | trigger

Target: metadata

Sharding factor: 1 (1 workers).

Comparing page_index_cache (976bb35) to bad29dd (merge-base) diff

Run configuration
run benchmark metadata
shards: 1

Results will be posted when all workers finish.


File an issue against this benchmark runner

@etseidl
etseidl marked this pull request as ready for review October 31, 2026 04:27
@adriangbot

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Target: metadata

Comparing page_index_cache (976bb35) to bad29dd (merge-base) diff

Run configuration
run benchmark metadata
shards: 1
Details
group                                               base                                   changed
-----                                               ----                                   -------
decode metadata (wide) with schema                  1.00     26.0±0.42ms        ? ?/sec    1.00     25.9±0.41ms        ? ?/sec
decode metadata (wide) with skip PES                1.00     26.4±0.39ms        ? ?/sec    1.00     26.3±0.39ms        ? ?/sec
decode metadata (wide) with skip all stats          1.00     30.1±0.31ms        ? ?/sec    1.00     30.1±0.31ms        ? ?/sec
decode metadata (wide) with skip column stats       1.01     27.8±0.30ms        ? ?/sec    1.00     27.5±0.28ms        ? ?/sec
decode metadata (wide) with skip size stats         1.00     30.4±0.24ms        ? ?/sec    1.00     30.3±0.26ms        ? ?/sec
decode metadata (wide) with stats mask              1.00     25.4±0.30ms        ? ?/sec    1.00     25.5±0.29ms        ? ?/sec
decode metadata with schema                         1.00      4.1±0.01µs        ? ?/sec    1.00      4.0±0.01µs        ? ?/sec
decode metadata with skip PES                       1.00      7.0±0.10µs        ? ?/sec    1.01      7.0±0.12µs        ? ?/sec
decode metadata with skip column stats              1.00      6.8±0.01µs        ? ?/sec    1.00      6.8±0.14µs        ? ?/sec
decode metadata with stats mask                     1.00      6.9±0.12µs        ? ?/sec    1.00      6.9±0.09µs        ? ?/sec
decode parquet metadata                             1.01      7.1±0.02µs        ? ?/sec    1.00      7.1±0.02µs        ? ?/sec
decode parquet metadata (wide)                      1.00     28.1±0.36ms        ? ?/sec    1.00     28.1±0.43ms        ? ?/sec
decode parquet metadata no path_in_schema (wide)    1.00     27.4±0.34ms        ? ?/sec    1.00     27.3±0.39ms        ? ?/sec
decode parquet metadata w/ size stats (wide)        1.01     34.1±0.32ms        ? ?/sec    1.00     33.9±0.35ms        ? ?/sec
open(default)                                       1.01      7.6±0.09µs        ? ?/sec    1.00      7.5±0.09µs        ? ?/sec
open(page index)                                    1.00    130.3±0.15µs        ? ?/sec    1.01    131.5±0.14µs        ? ?/sec
read page index                                     1.05   130.0±44.47µs        ? ?/sec    1.00    123.6±0.14µs        ? ?/sec
read page index reduced columns                     1.06     14.1±5.78µs        ? ?/sec    1.00     13.3±0.01µs        ? ?/sec

Per-runner information
metadata — shard 1/1

Node: gk3-benchmark-cluster-nap-8drekkdo-ff4449e6-kd9s

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

uname:

Linux bench-c6538969000-4206-9nzcd 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

BENCH_COMMAND:

cargo bench --features=arrow,async,test_common,experimental,object_store --bench metadata
CPU Details (lscpu)
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Resource Usage

base (merge-base)

Metric Value
Wall time 175.0s
Peak memory 148.8 MiB
Avg memory 64.8 MiB
CPU user 168.7s
CPU sys 0.1s
Peak spill 0 B

branch

Metric Value
Wall time 170.0s
Peak memory 148.7 MiB
Avg memory 63.5 MiB
CPU user 166.2s
CPU sys 0.1s
Peak spill 0 B

File an issue against this benchmark runner

@adriangb

Copy link
Copy Markdown
Contributor

Looks like perf is flat (which is good). I guess we don't have a direct measure of memory consumption per benchmark.

@etseidl

etseidl commented Nov 3, 2026

Copy link
Copy Markdown
Contributor Author

Looks like perf is flat (which is good). I guess we don't have a direct measure of memory consumption per benchmark.

Thanks for bringing this up @adriangb! While trying to answer I found a pretty big issue with the PR.

Everything is roses for sparse indexes. I had my agent whip up a quick test where the populated index entries were kept fixed (10 entries), but the declared dimensions of the table increase.

Declared dimensions (RG × columns) Populated entries per kind Main retained index bytes PR retained index bytes Main median lifecycle PR median lifecycle
10 × 100 10 280,954 3,450 9.092 µs 5.354 µs
100 × 100 10 2,805,274 3,450 87.288 µs 5.409 µs
100 × 1,000 10 28,005,274 3,450 14.646 ms 5.472 µs
1,000 × 1,000 10 280,048,474 3,450 151.662 ms 5.451 µs

Pretty good stuff.

But then I asked the obvious question, what about fully populated indexes? A comparison of fully populated 1000 x 1000 showed a significant slowdown for the new form (memory use was nearly flat, but total lifecycle time goes from 500ms to 1740ms), with most of the extra time going to populating the hash map in the builder, and then sorting the entries in the freeze stage.

That got me thinking that for the normal case where we're building an index from scratch, we're already doing the traversal in row-major order, and we know the exact bounds (mod missing indexes) up front. I had my agent whip up a test of going straight to the boxed slice form (skipping the builder), and was able to almost reach parity.

Workload Main Original PR Direct-vector PR
Fully populated 1,000 × 1,000 0.979 s 2.487 s 1.084 s
Sparse 10 × 100, 10 entries/kind 20.9 us 13.7 us 9.8 us
Sparse 1,000 × 1,000, same population 205.8 ms 13.8 us 9.7 us

I'll polish that work up and push it soon.

@alamb

alamb commented Nov 4, 2026

Copy link
Copy Markdown
Contributor

Is this something that I should put on my review queue @etseidl ?

@etseidl

etseidl commented Nov 4, 2026

Copy link
Copy Markdown
Contributor Author

Is this something that I should put on my review queue @etseidl ?

If you would be so kind. 🙏 otherwise I'll just keep making changes 😂

@etseidl

etseidl commented Nov 4, 2026

Copy link
Copy Markdown
Contributor Author

Seriously, though, it is fairly large at this point. If you'd rather concentrate on getting 60.2.0 shipped then I'm ok with waiting.

@alamb

alamb commented Nov 5, 2026

Copy link
Copy Markdown
Contributor

Thanks -- I'll try and give it a review.

BTW I highlighted the work to make metadata work better for metadata in a talk I am preparing for next weeks meetup if you are interested: https://docs.google.com/presentation/d/1Mlt4Ri9AjzwbJTB_xpfw3WU8nlDrh5i5FfF-U3cICkQ

@alamb

alamb commented Nov 5, 2026

Copy link
Copy Markdown
Contributor

Today I need to focus on getting DataFusion 56 release blocking issues ready

@alamb

alamb commented Nov 12, 2026

Copy link
Copy Markdown
Contributor

Sorry -- checking this one out again

@alamb alamb left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is great. Thank you @etseidl

I took the liberty of merging up from main and pushing some documentation tweaks to make the example easier to find.

validate_dimensions(self.num_row_groups, self.num_columns)?;
let num_entries = self.entries.len();
let mut entries = Vec::new();
entries.try_reserve_exact(num_entries).map_err(|err| {

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't understand the rationale for checking for an allocation failure -- we don't do that in the rest of the crate, do we? I think it is fine, I just found it kind of strange

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah, it's a mania my agent has picked up regarding fallibility. I'm inclined to keep it as a defense against future issues claiming some malformed input can overflow blah blah blah, but if you'd prefer we keep allocations consistent I can remove it (and the one right above it).

}

#[derive(Debug, Clone, PartialEq)]
pub(crate) struct ColumnChunkEntries<T> {

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This seems like it is very similar to the Map above, except that it has a single contiguous allocation -- what is the ratinale for the difference? Is it to reduce the in memory / heap size? If so perhaps we can document that rationale

Comment on lines +578 to +583
/// # Panics
///
/// Panics if either dimension exceeds `i32::MAX`, the maximum collection size representable
/// by Thrift.
///
/// Use [`try_new`](Self::try_new) for a fallible alternative.

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

allocating more than i32::MAX entries does sound like a bad idea to me and not a case that is important

.parse_and_finish(&file)
.unwrap();
let mut builder = metadata.into_builder();
let row_groups = builder

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I found this test hard to follow as the behavior to be validated (the row / columns) is interleaved with the mechnics of setting up the test

I would personally find it easier to read if there was some sort of human readable string that showed what the contents of the input row groups and the output row groups

assert!(metadata.page_index().is_some());
#[cfg(not(feature = "encryption"))]
assert_eq!(metadata.memory_size(), 12088);
assert_eq!(metadata.memory_size(), 9656);

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

that certainly looks a lot nicer 🎉

assert!(metadata.page_index().is_some());
#[cfg(not(feature = "encryption"))]
assert_eq!(metadata.memory_size(), 13929);
assert_eq!(metadata.memory_size(), 13993);

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this is slightly larger because there is some extra overhead per entry, right?

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah, there's a tradeoff here. We get better sparse memory usage, at the cost of a bit of overhead on storing all the keys inline. Thankfully that's offset some by dropping the Option wrapper on everything as well.

Down the road we could look at different storage for a fully populated index if this becomes and issue.

@etseidl

etseidl commented Nov 12, 2026

Copy link
Copy Markdown
Contributor Author

Thanks for the review @alamb. Please take one more quick look-see and let me know if you want any other changes before merging. 🙏

@alamb alamb left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good to me -- thank you @etseidl

this falls into the category of things that many people will benefit from / appreciate, but probably not even realize it is helping them

@alamb
alamb added this pull request to the merge queue Nov 13, 2026
@etseidl

etseidl commented Nov 13, 2026

Copy link
Copy Markdown
Contributor Author

Thanks @alamb! On to Bloom filters! 🚀 😄

this falls into the category of things that many people will benefit from / appreciate, but probably not even realize it is helping them

Next, on "Unsung Heroes of Parquet"... 😂

Merged via the queue into apache:main with commit 9cd60eb Nov 13, 2026
20 checks passed
@alamb

alamb commented Nov 13, 2026

Copy link
Copy Markdown
Contributor

Thanks @alamb! On to Bloom filters! 🚀 😄

this falls into the category of things that many people will benefit from / appreciate, but probably not even realize it is helping them

Next, on "Unsung Heroes of Parquet"... 😂

Well, we could make a Bay Area Parquet Meetup to tell the stories :) I am not sure I would necessarily watch that on youtube, but I would watch in real life

@etseidl
etseidl deleted the page_index_cache branch November 13, 2026 21:00
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

enhancement New feature/addition parquet Changes to the parquet crate performance

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Support load page index metadata with specified column id instead of all columns

4 participants