Repo refactor: rename "website" to "docs"

This commit is contained in:
Zef Hemel
2026-06-23 09:12:24 +02:00
parent 1b3e86d960
commit e7ccdc2883
226 changed files with 1 additions and 1 deletions
+317
View File
@@ -0,0 +1,317 @@
---
description: A Lua-embedded query syntax for selecting and transforming objects.
tags: glossary
---
Space Lua Integrated Query (SLIQ) is a SilverBullet specific Lua extension. It adds a convenient query syntax to the language in a backwards compatible way. It does so by overloading Lua’s default function call + single argument syntax when using `query` as the function call. As a result, Lua programs using SLIQ are still syntactically valid Lua.
The syntax for SLIQ is `query[[my query]]`. In regular Lua `[[my query]]` is just another way of writing `"my query"` (it is an alternative string syntax). Function calls that only take a string argument can omit parentheses, therefore `query[[my query]]` is equivalent to `query("my query")`.
However, in [[Space Lua]] it is interpreted as an SQL- and [LINQ](https://learn.microsoft.com/en-us/dotnet/csharp/linq/)-inspired integrated query language.
General syntax:
```postgres
query [[
from <<expression>>
[ where <<expression>> ]
[ group by <<expression>> [, ...] ]
[ having <<expression>> ]
[ order by <<sort_expression>> [, ...] ]
[ limit <<expression>> [, <<expression>>] ]
[ offset <<expression>> ]
[ select <<expression>> ]
]]
```
Unlike in SQL where clauses must be written in a particular order, SLIQ clauses can be written in any order, and the only mandatory clause is the `from` clause.
The `order by` clause uses `<<sort_expression>>`:
```postgres
<<expression>>
[ asc | desc | using <<comparator>> ]
[ nulls { first | last } ]
```
If an aggregate function call is used for projection (`select`), the general syntax is:
```postgres
<<aggregator>>( <<expression>> [, ...]
[ order by <<sort_expression>> [, ...] ]
[ filter (where <<expression>>) ]
)
```
See [[Space Lua/Integrated Query/Aggregating]] for details.
Aggregate functions support an optional intra-aggregate `order by` and/or `filter` clause:
<aggregate>(<expression> [order by <expr> [asc|desc] [nulls {first|last}], ...])
<aggregate>(<expression> ...) filter(where <condition>)
These can be combined. See [[Space Lua/Integrated Query/Aggregating]] for details.
SLIQ operates on any Lua collection.
For instance, to sort a list of numbers in descending order:
${query[[from n = {1, 2, 3} order by n desc]]}
However, in most cases you’ll use it in conjunction with [[API/index#index.objects(tag)]] (or one of its convenience wrappers like `index.pages()`). Here’s an example querying the 3 pages that were last modified:
${query[[
from p = index.pages()
order by p.lastModified desc
select p.name
limit 3
]]}
Note that the query returns a regular Lua table, so it can be part of a bigger expression:
${some(query[[
from p = index.pages()
limit 0
]]) or "Matched no pages"}
# Clauses
Here are the clauses that are currently supported:
## `from`
The `from` clause specifies the source of your data. There are two syntactic variants:
**Recommended:** With explicit variable binding:
from v = <<expression>>
binding each item to the variable `v`.
However, there is also the more concise:
from <<expression>>
implicitly binding each item to the variable `_` as well as making all attributes directly available as variables. The latter, while shorter, is less performant and will block future optimizations, so the variable-binding variant is preferred.
> **warning** Warning
> When you use a `from` clause without explicit variable binding (so without the `v in` syntax), note that any attribute of the object you’re iterating over will shadow global variables. For instance, if you have an object with a `table` attribute, regular `table` APIs will become inaccessible within the query.
>
> **Recommendation:** Use the explicit variable binding syntax
Example without variable binding:
${query[[from {1, 2, 3} select _]]}
With variable binding:
${query[[from n = {1, 2, 3} select n]]}
A more realistic example using `index.pages`:
${query[[from p = index.pages() order by p.lastModified select p.name limit 3]]}
## `where`
The `where` clause allows you to filter data. When the expression evaluated to a truthy value, the item is included in the result.
Example:
${query[[from n = {1, 2, 3, 4, 5} where n > 2]]}
Or to select 5 pages tagged with `#meta`:
${query[[from p = index.pages() where table.includes(p.tags, "meta") limit 5]]}
Or select based on name (including folder) and a [[API/string|string function]]:
${query[[from p = index.pages() where p.name:startsWith("Person")]]}
## `group by`
The `group by` clause groups results by one or more key expressions. After grouping, each result row becomes a table with two fields:
- `key` — the group key value (single value for one key, table for multi-key)
- `group` — a table (array) of all original items in that group
The `group by` field names are also available as bare variables in `having`, `select`, and `order by`. Use `#group` to get the count of items in a group.
Example:
${query[[
from p = index.objects("tag")
group by p.name
select { name = name, count = #group }
limit 5
]]}
See [[Space Lua/Integrated Query/Grouping]] for detailed examples.
## `having`
The `having` clause filters groups **after** `group by`. It follows SQL semantics: only group key fields, `key`, and `group` are accessible — use `where` to filter individual rows before grouping.
Aggregate functions like `count()`, `sum()`, `min()`, `max()`, and `avg()` can be used in `having` expressions. See [[Space Lua/Integrated Query/Aggregating]] for details.
Example:
${query[[
from p = index.tags()
group by p.name
having #group > 2
select { name = name, count = #group }
order by count desc
limit 5
]]}
See [[Space Lua/Integrated Query/Grouping]] for detailed examples.
## `order by`
The `order by` clause sorts results by one or more expressions. By default, sorting is ascending. Append `desc` for descending order, or `asc` to be explicit about ascending.
As an example, the last 3 modified pages:
${query[[
from p = index.pages()
order by p.lastModified desc
select p.name
limit 3
]]}
You can sort by multiple expressions separated by commas. Each key is evaluated left to right — the second key only matters when the first compares as equal:
${query[[
from p = index.pages()
order by p.lastModified desc, p.name
select p.name
limit 3
]]}
Each sort key can have its own direction:
```lua
query[[
from p = data
order by p.category asc, p.priority desc
select { name = p.name }
]]
```
### Null placement
By default, `nil` values follow SQL conventions: they appear **last** for ascending order and **first** for descending order. You can override this per key with `nulls first` or `nulls last`:
${query[[
from p = index.pages()
order by p.priority desc nulls last
select { name = p.name, priority = p.priority }
limit 3
]]}
### String collation
Sorting of strings can be adjusted with `queryCollation` in [[^Library/Std/Config]].
### `using` (custom comparators)
The `using` clause specifies a custom comparator function instead of the default `asc`/`desc` ordering. The two are mutually exclusive — `using` defines both the comparison logic and the direction.
The comparator must accept two arguments and return `true` when the first should come strictly before the second. It can be a named function:
```lua
function byLength(a, b)
return #a < #b
end
```
```lua
query [[
from p = index.pages()
order by p.name using byLength
select p.name
limit 5
]]
```
Or an anonymous function inline. Example:
${query[[
from n = {5, 1, 3, 2, 4}
order by n using function(a, b) return a < b end
]]}
The `nulls` clause works with `using`, and each sort key can independently choose `asc`/`desc` or `using`:
```lua
query [[
from p = data
order by
p.category using customCategoryCmp,
p.priority desc nulls last
select {
name = p.name
}
]]
```
When `using` is specified, it overrides any `queryCollation` configuration for that sort key.
> **note** Note
> `using` is a reserved keyword in Space Lua and cannot be used as a variable name.
#### Strict weak ordering
A comparator must satisfy **strict weak ordering** (SWO) — if comparing A with B returns `true`, then comparing B with A must return `false`. In practice this means using strict comparisons like `<` or `>` and **never** `<=` or `>=`.
The query engine validates this at runtime. If comparing two values in both directions both return `true`, the query fails with a clear error:
${query [[
from n = {5, 1, 3, 2, 3}
order by n using function(a, b) return a <= b end
]]}
The query engine uses a *stable merge sort* algorithm with guaranteed performance. Items that compare as equal preserve their original order and an invalid comparator cannot cause an infinite loop or crash — the violation is detected and reported as an error.
## `limit`
The `limit` clause allows you to limit the number of results, optionally with an inline offset.
Example:
${query[[from {1, 2, 3, 4, 5} limit 3]]}
You can also specify an offset as a second argument to skip some results:
${query[[from {1, 2, 3, 4, 5} limit 3, 2]]}
> **note** Note
> If both an inline offset (`limit 3, 2`) and a standalone `offset` clause are present, the last one encountered wins.
## `offset`
The `offset` clause skips a number of rows from the beginning of the result set. It can be used with or without `limit`.
Skip the first 2 results:
${query[[from {1, 2, 3, 4, 5} offset 2]]}
Combined with `limit`:
${query[[from {1, 2, 3, 4, 5} offset 2 limit 3]]}
If the offset is larger than the number of available rows, the result is empty — this is not an error, matching PostgreSQL semantics.
> **note** Note
> If both a standalone `offset` clause and an inline offset in `limit` (`limit 3, 2`) are present, the last one encountered wins.
## `select`
The `select` clause allows you to transform each item in the result set. If omitted, it defaults to returning the item itself.
When used with `group by`, aggregate functions like `sum()`, `count()`, `min()`, `max()`, `avg()`, and `array_agg()` can be used in the `select` expression to compute values across each group. Aggregates also support intra-aggregate `order by` to control the order in which values are processed, and `filter(where ...)` to restrict which rows contribute. See [[Space Lua/Integrated Query/Aggregating]] for details.
Some examples:
Double each number:
${query[[from n = {1, 2, 3} select n * 2]]}
It is convenient to combine it with the [[API/table#table.select(table, keys...)]] API:
${query[[
from p = index.pages()
select table.select(p, "name", "lastModified")
limit 3
]]}
# Rendering the output
To render the output as a template, you can rely on the fact that queries return Lua tables. For example, to apply a template to render every page as a link:
${query[[
from p = index.pages()
order by p.lastModified desc
limit 3
select templates.pageItem(p)
]]}
To render pages as links with their full local URL, use `templates.fullPageItem`. For more information on available templates, see [[^Library/Std/Infrastructure/Query Templates]].