data explorer: set edges to value of single bin (#884)

* set edges to value of single bin

* update changed behaviour

Author: isabelizimm

crates/ark/src/data_explorer/histogram.rs

@@ -379,8 +379,8 @@ mod tests {
     fn test_constant_column() {
         r_task(|| {
             // This is the default `hist` behavior, single bin containing all info.
-            test_histogram("c(1, 1, 1)", 4, vec!["0.00", "1.00"], vec![3]);
-            test_histogram_method("c(1, 1, 1)", "sturges", vec!["0.00", "1.00"], vec![3])
+            test_histogram("c(1, 1, 1)", 4, vec!["1.00", "1.00"], vec![3]);
+            test_histogram_method("c(1, 1, 1)", "sturges", vec!["1.00", "1.00"], vec![3])
         })
     }
 

crates/ark/src/modules/positron/r_data_explorer.R

@@ -473,6 +473,17 @@ profile_histogram <- function(
     bin_edges <- h$breaks
     bin_counts <- h$counts
 
+    # Special case: if we have a single bin, check if all values are the same
+    # If so, override the bin edges to be the same value instead of value +/- 1
+    if (length(bin_counts) == 1 && length(x) > 0) {
+        # Check if all values are the same
+        unique_values <- unique(x)
+        if (length(unique_values) == 1) {
+            # All values are the same, set bin edges to [value, value]
+            bin_edges <- c(unique_values[1], unique_values[1])
+        }
+    }
+
     # For dates, we convert back the breaks to the date representation.
     if (inherits(x, "POSIXct")) {
         # Must supply an `origin` on R <= 4.2

crates/ark/tests/data_explorer.rs

@@ -1986,6 +1986,51 @@ fn test_histogram() {
     });
 }
 
+#[test]
+fn test_histogram_single_bin_same_values() {
+    let _lock = r_test_lock();
+
+    let socket = open_data_explorer_from_expression("data.frame(x = rep(5, 10))", None).unwrap();
+
+    let make_histogram_req = |id, column_index, method, num_bins, quantiles| {
+        DataExplorerBackendRequest::GetColumnProfiles(GetColumnProfilesParams {
+            callback_id: id,
+            profiles: vec![ColumnProfileRequest {
+                column_index,
+                profiles: vec![ColumnProfileSpec {
+                    profile_type: ColumnProfileType::SmallHistogram,
+                    params: Some(ColumnProfileParams::SmallHistogram(ColumnHistogramParams {
+                        method,
+                        num_bins,
+                        quantiles,
+                    })),
+                }],
+            }],
+            format_options: default_format_options(),
+        })
+    };
+
+    let id = String::from("histogram_same_values");
+    let req = make_histogram_req(id.clone(), 0, ColumnHistogramParamsMethod::Fixed, 5, None);
+
+    expect_column_profile_results(&socket, req, |profiles| {
+        let histogram = profiles[0].small_histogram.clone().unwrap();
+
+        // When all values are the same, we should get a single bin with count = number of values
+        assert_eq!(histogram.bin_counts, vec![10]);
+
+        // The bin edges should be [5, 5] since all values are 5
+        let expected_edges = r_task(|| {
+            format_string(
+                harp::parse_eval_global("c(5, 5)").unwrap().sexp,
+                &default_format_options(),
+            )
+        });
+        assert_eq!(histogram.bin_edges, expected_edges);
+        assert_eq!(histogram.quantiles, vec![]);
+    });
+}
+
 #[test]
 fn test_frequency_table() {
     let _lock = r_test_lock();