Dev

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "tsvkit"
-version = "0.9.3"
+version = "0.9.5"
 edition = "2024"
 
 [dependencies]

README.md

@@ -148,7 +148,7 @@ The same expression language powers `filter -e`, `mutate -e name=EXPR`, and rege
 | `+ - * / ^` | Arithmetic operators (`^` is exponentiation, right-associative). | Numbers |
 | `== != < <= > >=` | Comparisons. | Numbers or strings |
 | `&` / `and` | Logical AND. | Booleans |
-| `|` / `or` | Logical OR. | Booleans |
+| `\|` / `or` | Logical OR. | Booleans |
 | `!` / `not` | Logical negation. | Booleans |
 | `~` | Regex match. Right-hand side can be literal text or a `$range`. | Strings |
 | `!~` | Regex does *not* match. | Strings |
@@ -165,9 +165,27 @@ The same expression language powers `filter -e`, `mutate -e name=EXPR`, and rege
 | `ln(expr)` | Natural logarithm |
 | `log(expr)` / `log10(expr)` | Base-10 logarithm |
 | `log2(expr)` | Base-2 logarithm |
+| `len(expr)` | Character count using Unicode code points. |
+| `is_na(expr)` | Returns `1` when the expression is blank/`NA`/`NaN`, otherwise `0`. |
 
 Functions accept column references (`abs($purity - 1)`), constants, or subexpressions. Empty or non-numeric values yield blanks.
 
+**Conditional and regex helpers**
+
+- `case_when(condition -> result, ..., _ -> default)` evaluates each boolean condition in order and returns the matching result. The final `_` branch acts as the default.
+- `switch(value, [match1, match2] -> result, ..., _ -> default)` compares `value` to one or more literal matches (strings or numbers) and returns the corresponding result.
+- `re(value, pattern)` evaluates a regex against `value`, returning `1` or `0`. When the pattern matches, capture groups become available as `$1`, `$2`, etc. for the remainder of the expression (use `$0`-style numeric selectors sparingly when you rely on captures).
+
+Example:
+
+```
+case_when(
+    re($sample, "^ERR(\\d+)$") -> $1,
+    re($sample, "^SRR")          -> "SRA",
+    _                              -> $sample
+)
+```
+
 **Row-wise aggregation helpers**
 
 Available within `mutate` expressions via functions such as `sum($col1:$col5)`; see the [Mutate](#mutate) section for the full list.
@@ -225,7 +243,7 @@ tsvkit filter -e '$group == "case" & $purity >= 0.94' examples/samples.tsv
 | Literals | `1.25`, `"case"` | Strings use double quotes; escape inner quotes with `\"`. |
 | Arithmetic | `($rna_ug - $dna_ug) / $rna_ug` | Standard precedence applies (parentheses for clarity). |
 | Comparisons | `$purity >= 0.9`, `$group != "control"` | Works on numeric or string data. |
-| Logical | `($purity >= 0.9) & ($group == "case")` | `&`, `|`, and `!` (or `and`, `or`, `not`). |
+| Logical | `($purity >= 0.9) & ($group == "case")` | `&`, `\|`, and `!` (or `and`, `or`, `not`). |
 | Numeric functions | `log2($total)`, `sqrt($reads)` | See [Expression language essentials](#expression-language-essentials). |
 | Row-wise aggregators | `sum($dna_ug:$rna_ug)`, `mode($1,$3)`, `countunique($gene:)` | Same catalog as [`summarize`](#summarize): totals, quantiles (`q*` / `p*`), variance/SD, products, entropy, argmin/argmax, membership stats. Works with ranges, lists, and open selectors. |
 | Regex match | `$tech ~ "sRNA"`, `$notes !~ "(?i)fail"` | Patterns follow Rust `regex` syntax. `(?i)` enables case-insensitive matching. |
@@ -260,6 +278,19 @@ tsvkit mutate \
   examples/cytokines.tsv
 ```
 
+Use `case_when`, `switch`, and the `re()` helper for richer branching logic and regex capture reuse:
+
+```bash
+tsvkit mutate \
+  -e 'label = case_when(
+        re($sample, "^ERR(\d+)$") -> $1,
+        re($sample, "^SRR")        -> "SRA",
+        _                            -> $sample
+      )' \
+  -e 'bucket = case_when(len($clean) == 0 -> "empty", len($clean) < 5 -> "short", _ -> "long")' \
+  examples/samples.tsv
+```
+
 Apply in-place edits with the sed-style form:
 
 ```bash

src/common.rs

@@ -7,12 +7,28 @@ use csv::ReaderBuilder;
 use flate2::read::MultiGzDecoder;
 use xz2::read::XzDecoder;
 
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub enum SpecialColumn {
+    FilePath,
+    FileBase,
+}
+
+impl SpecialColumn {
+    pub fn default_header(self) -> &'static str {
+        match self {
+            SpecialColumn::FilePath => "__file__",
+            SpecialColumn::FileBase => "__base__",
+        }
+    }
+}
+
 #[derive(Debug, Clone)]
 pub enum ColumnSelector {
     Index(usize),
     FromEnd(usize),
     Name(String),
     Range(Option<Box<ColumnSelector>>, Option<Box<ColumnSelector>>),
+    Special(SpecialColumn),
 }
 
 pub fn parse_selector_list(spec: &str) -> Result<Vec<ColumnSelector>> {
@@ -86,6 +102,12 @@ pub fn resolve_selectors(
     let mut indices = Vec::with_capacity(selectors.len());
     for selector in selectors {
         match selector {
+            ColumnSelector::Special(special) => {
+                bail!(
+                    "special column '{}' cannot be resolved as a positional index",
+                    special.default_header()
+                );
+            }
             ColumnSelector::Index(_) | ColumnSelector::FromEnd(_) | ColumnSelector::Name(_) => {
                 let index = resolve_selector_index(headers, selector, no_header)?;
                 indices.push(index);
@@ -207,6 +229,21 @@ pub fn should_skip_record(
     false
 }
 
+pub fn inconsistent_width_error(
+    source: &str,
+    row_number: usize,
+    expected: usize,
+    actual: usize,
+) -> anyhow::Error {
+    anyhow!(
+        "rows in {} have inconsistent column counts at row {} (expected {}, got {})",
+        source,
+        row_number,
+        expected,
+        actual,
+    )
+}
+
 pub fn default_headers(len: usize) -> Vec<String> {
     (1..=len).map(|i| format!("col{}", i)).collect()
 }
@@ -251,6 +288,11 @@ fn parse_simple_selector(token: &str) -> Result<ColumnSelector> {
     if let Some(literal) = parse_brace_literal(token)? {
         return Ok(ColumnSelector::Name(literal));
     }
+    match token {
+        "__file__" => return Ok(ColumnSelector::Special(SpecialColumn::FilePath)),
+        "__base__" => return Ok(ColumnSelector::Special(SpecialColumn::FileBase)),
+        _ => {}
+    }
     if let Some(stripped) = token.strip_prefix('-') {
         if stripped.is_empty() {
             bail!("column selector '-' must include an index");
@@ -520,6 +562,10 @@ fn resolve_selector_index(
                 .with_context(|| format!("column '{}' not found", name))?;
             Ok(index)
         }
+        ColumnSelector::Special(special) => bail!(
+            "special column '{}' not supported without column injection",
+            special.default_header()
+        ),
         ColumnSelector::Range(_, _) => {
             bail!("unexpected nested column range")
         }
@@ -528,7 +574,10 @@ fn resolve_selector_index(
 
 #[cfg(test)]
 mod tests {
-    use super::{ColumnSelector, parse_selector_list, parse_single_selector, resolve_selectors};
+    use super::{
+        ColumnSelector, SpecialColumn, parse_selector_list, parse_single_selector,
+        resolve_selectors,
+    };
 
     #[test]
     fn resolves_name_range() {
@@ -642,4 +691,15 @@ mod tests {
         let err = parse_selector_list("`foo").unwrap_err();
         assert!(err.to_string().contains("unterminated backtick"));
     }
+
+    #[test]
+    fn distinguishes_injected_and_literal_file_columns() {
+        let selectors = parse_selector_list("__file__,{__file__},`__base__`").unwrap();
+        assert!(matches!(
+            selectors[0],
+            ColumnSelector::Special(SpecialColumn::FilePath)
+        ));
+        assert!(matches!(selectors[1], ColumnSelector::Name(ref name) if name == "__file__"));
+        assert!(matches!(selectors[2], ColumnSelector::Name(ref name) if name == "__base__"));
+    }
 }