feat(api): add argmax and argmin column methods

ibis/expr/operations/reductions.py

@@ -191,6 +191,20 @@ class Min(Filterable, Reduction):
     output_dtype = rlz.dtype_like('arg')
 
 
+@public
+class ArgMax(Filterable, Reduction):
+    arg = rlz.column(rlz.any)
+    key = rlz.column(rlz.any)
+    output_dtype = rlz.dtype_like("arg")
+
+
+@public
+class ArgMin(Filterable, Reduction):
+    arg = rlz.column(rlz.any)
+    key = rlz.column(rlz.any)
+    output_dtype = rlz.dtype_like("arg")
+
+
 @public
 class ApproxCountDistinct(Filterable, Reduction):
     """Approximate number of unique values using HyperLogLog algorithm.

ibis/expr/types/generic.py

@@ -652,15 +652,33 @@ def approx_median(
         return ops.ApproxMedian(self, where).to_expr().name("approx_median")
 
     def max(self, where: ir.BooleanValue | None = None) -> Scalar:
+        """Return the maximum of a column."""
         import ibis.expr.operations as ops
 
         return ops.Max(self, where).to_expr().name("max")
 
     def min(self, where: ir.BooleanValue | None = None) -> Scalar:
+        """Return the minimum of a column."""
         import ibis.expr.operations as ops
 
         return ops.Min(self, where).to_expr().name("min")
 
+    def argmax(
+        self, key: ir.Value, where: ir.BooleanValue | None = None
+    ) -> Scalar:
+        """Return the value of `self` that maximizes `key`."""
+        import ibis.expr.operations as ops
+
+        return ops.ArgMax(self, key=key, where=where).to_expr()
+
+    def argmin(
+        self, key: ir.Value, where: ir.BooleanValue | None = None
+    ) -> Scalar:
+        """Return the value of `self` that minimizes `key`."""
+        import ibis.expr.operations as ops
+
+        return ops.ArgMin(self, key=key, where=where).to_expr()
+
     def nunique(
         self, where: ir.BooleanValue | None = None
     ) -> ir.IntegerScalar: