[PGPRO-12159] Added functions for exploring the pages of the rum index. #150
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
arseny114
changed the title
arseny114
force-pushed
the
PGPRO-12159
branch
3 times, most recently
from
d13ae48 to
63ecfd5
Compare
arseny114
force-pushed
the
PGPRO-12159
branch
6 times, most recently
from
e1ddf10 to
94bf9c1
Compare
arseny114
force-pushed
the
PGPRO-12159
branch
4 times, most recently
from
f6af174 to
fd282ce
Compare
arseny114
force-pushed
the
PGPRO-12159
branch
6 times, most recently
from
af94bb0 to
ec20040
Compare
This commit adds three functions for low-level exploration
of the index's rum pages:
1) rum_metapage_info() -- is used to examine the information
posted on the meta page (flags: {meta}).
2) rum_page_opaque_info() -- is used to examine information
that is placed in the opaque area of the index page (any
index page).
3) rum_leaf_data_page_items() -- is used to examine the
information that is placed on the leaf pages of the
posting tree (flags: {leaf, data}).
To extract information, all these functions need to pass
the index name and the page number.
Tags: rum
1) rum_internal_data_page_items() - it is intended for viewing information
that is located on the internal pages of the posting tree (flags {data}).
2) rum_leaf_entry_page_items() - it is intended for viewing information
that is located on the leaf pages of the entry tree (flags {leaf}).
3) rum_internal_entry_page_items() - it is intended for viewing information
that is located on the internal pages of the entry tree (flags {}).
To extract information, all these functions need to pass the index name
and the page number.
Tags: rum
If you create an index with the operator class rum_tsvector_ops, the positions of the lexemes will be saved as additional information. The positions are stored in compressed form in bytea. There is a problem that is related to the fact that in the posting tree, additional information for the senior keys is stored in a different way, which is why it has not yet been possible to output it. For all other cases, the output of additional information works correctly. Tags: rum
If the index is created with the appropriate class of operators, then in addition to the positions of the lexemes, weights (A, B, C, D) are also stored in the additional information. Their output has been added. In addition, Asserts have been added to the find_add_info_atr_num() and find_add_info_oid() functions, which check that there is only one (or zero) type of additional information in the index. Tags: rum
The crashes were due to the fact that the construct_array_builtin() function is not defined on versions below 16. Tags: rum
The crashes were due to the fact that the errdetail_relkind_not_supported() function is not defined on versions below 15. Tags: rum
The crashes were due to the fact that the MONEYOID is not defined on versions below 14. Tags: rum
Added functions for checks. Added auxiliary functions to reduce code duplication. Tags: rum
Added an auxiliary universal function for scanning leaf and internal pages of the Posting Tree. Added an auxiliary function for reading the high key from the pages of the Posting Tree. Tags: rum
added 9 commits
October 7, 2025 18:35
rum_debug_funcs tests added only for enterprise. Tags: rum
Refactoring has been performed. The code is divided into more logical blocks using added functions and macros with clear names. Tags: rum
Previously, it was assumed that the RUM index could not contain attributes with different types of additional information (and some functions from PGPRO-12159 relied on this), but it is possible. If you create an index for two columns of the tsvector type and attach another column to one of them, it turns out that one tsvector has the attach attribute as additional information, and the second one will have positions. In order to fix this, the functions that find the attribute number of the key for which the posting tree was built were redesigned. Now they scan the index and look for the key (and its attribute number) for the scanned page. Tags: rum
Tags: rum
Tags: rum
The regression tests have been removed. Tags: rum
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.
Six functions have been developed to extract information from the pages of the RUM index:
To extract information, all these functions need to pass the index name and the page number.
Tags: rum