Version 5.9.1¶

Released on 2024-10-28.

Note

If you are upgrading a cluster, you must be running CrateDB 4.0.2 or higher before you upgrade to 5.9.1.

We recommend that you upgrade to the latest 5.8 release before moving to 5.9.1.

Before upgrading, you should back up your data.

Warning

Due to a BWC issue, a rolling upgrade from 5.8.x to 5.9.1 can lead to problems when running queries as a user created on an older version during a rolling upgrade.

Please do a rolling upgrade from 5.8.x to 5.9.2 or higher instead.

Warning

Tables that were created before CrateDB 4.x will not function with 5.x and must be recreated before moving to 5.x.x.

You can recreate tables using COPY TO and COPY FROM or by inserting the data into a new table.

Table of contents

See the Version 5.9.0 release notes for a full list of changes in the 5.9 series.

Fixes ¶

Fixed an issue which would allow to cast a number to a NUMERIC, even if the NUMERIC’s’ precision and scale would not be sufficient to fit the number without loosing precision, instead of throwing an error, e.g.:
```
SELECT CAST(123.45 as NUMERIC(4, 2))
```
This fix also changes the behavior, to throw an error when attempting to insert such values to a NUMERIC table column, instead of successfully inserting them and loose their precision, e.g.:
```
CREATE TABLE tbl (a NUMERIC(4,2));
INSERT INTO tbl(a) VALUES(123.456);
```

Fixed and issue that would cause error to be thrown when using correlated subquery with join on a primary key column, e.g.:

CREATE TABLE tbl(x int PRIMARY KEY);
SELECT (
    SELECT x FROM tbl a WHERE a.x = b.x LIMIT 1
) AS t
FROM (SELECT x FROM tbl) as b;

Fixed an issue that would cause an error to be thrown when using PostgreSQL wire protocol and attempting to insert 0 value for a numeric NUMERIC, or trying to get 0 numeric value from a query, e.g.:
```
SELECT 0::numeric;
```
Fixed behavior of comparing (=, >, etc. ) a column of type CHAR with string literals, so that any trailing whitespaces are ignored for both sides of the comparison, thus matching PostgreSQL behavior.
Fixed an issue that could cause comparisons (=, >, etc. ) of columns of CHAR type to yield different results when in the WHERE clause, than in any other part of a query, e.g. in the SELECT clause.

Fixed an issue that led to a NullPointerException when the same filter expression was used multiple times in an aggregation on an aliased table. For example:

SELECT
    alias1.country,
    MAX(height) FILTER (WHERE height>0) as max_height,
    MAX(prominence) FILTER (WHERE height>0) as max_prominence
FROM
    sys.summits alias1
GROUP BY
    alias1.country
LIMIT 100;

Fixed an issue that would cause users and roles to loose irrelevant privileges, like: Administration Language (AL), when a table or a view is dropped.
Improved the memory accounting for values of type geo_shape to avoid running into out of memory errors or higher than expected GC load when running queries like INSERT INTO ... (query).
Fixed a regression introduced in 5.8.4 leading to a rejection of writes into a column of the OBJECT(IGNORED) type if it had an array sub-column with mixed types.
Fixed an issue resulting in different (and partly wrong) sys.allocations.decisions results when querying the table on different nodes.
Fixed a BWC issue resulting in an exception when querying the sys.users table on an older node (< Version 5.9.0) during a rolling upgrade.
Fixed an incorrect optimization of comparison operators in combination with array subscript expression, ANY, ARRAY_UPPER and ARRAY_LENGTH function argument whereas the array (or array element) column has an explicit cast. Such queries may return wrong results as the explicit cast was removed. Example:

WHERE my_array[1]::timestamptz > 3::timestamptz
Fixed an issue that caused = comparison on NUMERIC types with unmatched precisions and scales to return invalid results or throw an IllegalStateException.
Fixed an issue that caused write operations to fail if a table contains a generated geo_shape column with type Polygon, MultiPolygon, LineString or MultiLineString and a user provides a correct value for this generated column.
Fixed an issue that caused >, <, >= or <= on NUMERIC types with unmatched precisions and scales or negative values to return invalid results.

Version 5.9.1¶

Fixes¶

Fixes ¶