1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
|
= New Features
* As Set will be a core class in Ruby 4+, this release integrates Set
support into Sequel. Dataset#as_set and #select_set have been added
to return sets. Dataset#as_set is similar to Dataset#map, and
Dataset#select_set is similar to Dataset#select_map, except that
Dataset#as_set and #select_set return sets instead of arrays.
Sets are now supported in dataset filters:
DB[:t].where(c: Set["d", "e"])
# SELECT * FROM t WHERE (c IN ('d', 'e'))
The empty_array_consider_nulls and split_array_nil extensions now
support sets in addition to arrays.
Dataset#qualify now supports sets.
Passing a set as an argument to Dataset#where, #exclude, and similar
filtering methods now raises an error.
Database typecasting of sets to string is now disallowed, similar to
how typecasting hashes and arrays to string is already disallowed.
Filtering by associations now supports a set of associated objects.
The constraint_validations extension now supports sets of integers
or strings for include validations.
The eval_inspect and to_dot extensions now supports sets.
The pg_auto_parameterize extension now auto parameterizes IN/NOT IN
filters with sets of integers. The pg_auto_parameterize_in_array now
auto parameterizes IN/NOT IN with sets of other values.
The pg_hstore_ops and pg_json_ops extensions now support sets in
situations where arrays are supported and order is not relevant.
The static_cache plugin now optimizes Model#as_set calls to avoid a
database query. The subset_static_cache plugin now optimizes
Dataset#as_set calls for static subsets to avoid a database query.
Dataset#full_text_search on Microsoft SQL Server, MySQL, and
PostgreSQL now supports a set for the terms argument.
* The create_table index method and related methods on PostgreSQL 11+
now support an :only option to create an index only for the specific
table, and not any child tables.
* Database#indexes on PostgreSQL now supports an :invalid option. To
include invalid indexes, you can use invalid: :include. To limit the
returned indexes to only invalid indexes, you can use
invalid: :only.
= Other Improvements
* Requiring sequel/model will now require sequel_pg/model if sequel_pg
has been previously required and sequel_pg 1.18+ is in use. This
allows for more optimizations in the case where sequel/core is
required, a connection to a PostgreSQL database is made (requiring
the postgres adapter and sequel_pg), and then sequel/model is
required. This is a common and recommended pattern so that code that
only needs access to a Sequel::Database and not Sequel models does
not need to load the model support, so it was considered worth
optimizing.
= Backwards Compatibility
* The set_literalizer extension has been deprecated. Most of what the
set_literalizer extension supports is now supported by default. The
one difference is the set_literalizer extension will treat a set
containing only arrays with two elements the same as an array
containing only arrays with two elements, and the default set
support deliberately does not do that.
|
