Expressions and types
Arithmetic, comparisons, boolean logic, case, casts, nullable columns and scalar bindings, strings, dates, and typed scalar / series / table values.
Reference
Ibex language reference
An overview of the language and runtime. Find the operation you need here, then follow the linked guides or specification for complete syntax and semantics.
At a glance
The shape of an Ibex program
Programs bind immutable values, then transform tables with named clauses. Clauses run from left to right inside a table expression.
import "csv";
let prices = csv::read("prices.csv");
let summary = prices[
filter volume > 0,
select { average = mean(price), traded = sum(volume) },
by symbol,
order { traded desc }
];Language
Core forms and table operations
let name = value | Bind an immutable scalar, series, table, or function result. |
fn name(args) -> Type { ... } | Declare a typed reusable function. |
import "plugin" | Load a bundled I/O plugin. |
extern fn ... from "header.hpp" | Declare a native function or source. |
extern type Name from "header.hpp" | Declare an opaque resource, such as a database connection, that functions open, pass, return, and close. See the connections guide. |
Table { field = values } | Create a table from column expressions or literals. |
filter predicate | Keep rows for which the predicate is true. |
select { fields } | Project fields, create computed fields, or aggregate with by. |
update { fields } | Add or replace fields while preserving the remaining columns. |
map { name = expr, ... } | Evaluate scalar fields once per input row and return exactly the named columns. Rows and fields run in order, so effectful extern calls are supported. |
by key | Group a select or update. |
order { keys } | Sort rows using ascending or descending keys. |
rename, distinct, head, tail | Rename, deduplicate, or limit rows. |
See the transformation tutorial for first use, and the language specification for grammar and exact rules.
Functionality
Find an operation by task
Aggregation and windows
Grouped reduction, grouped update, ranking, cumulative functions, rolling windows, resampling, and TimeFrames.
Joins and reshaping
Inner, outer, semi, anti, cross, and as-of joins; row and column binding; wide-to-long and long-to-wide reshaping.
Input and output
CSV, JSON, Parquet, directory enumeration, command-line argument parsing, SQLite/ADBC, Kafka, and cloud-backed paths through plugins.
Functions and metaprogramming
Typed user functions, native extern functions, compile-time map expansion, and table-schema introspection.
Interactive and host use
The REPL, script execution, C++23 transpilation, Python notebooks, and the R/dplyr integration.
Advanced workflows
C++ code generation, continuous streams, producer queues, and formula-based model specifications.