From: Ben Pfaff <blp@cs.stanford.edu>
Date: Sat, 21 Dec 2024 20:19:38 +0000 (-0800)
Subject: work on dictionary reading
X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=d3e95332d83571342ff3fbb104a1e572f905bc68;p=pspp

work on dictionary reading
---

diff --git a/rust/pspp/src/cooked.rs b/rust/pspp/src/cooked.rs
index 7bfba78ecb..4d46dc4ead 100644
--- a/rust/pspp/src/cooked.rs
+++ b/rust/pspp/src/cooked.rs
@@ -552,7 +552,7 @@ pub fn decode(
         }
 
         for dict_index in dict_indexes {
-            let mut variable = &dictionary.variables[dict_index];
+            let variable = dictionary.variables.get_index_mut2(dict_index).unwrap();
             for ValueLabel { value, label } in record.labels.iter().cloned() {
                 let value = match value {
                     raw::Value::Number(number) => Value::Number(number.map(|n| n.into())),
@@ -560,6 +560,7 @@ pub fn decode(
                         string.0[..variable.width.as_string_width().unwrap()].into()
                     }
                 };
+                variable.value_labels.insert(value, label);
             }
         }
     }
diff --git a/rust/pspp/src/dictionary.rs b/rust/pspp/src/dictionary.rs
index d4fdd6c458..e8dcec2f20 100644
--- a/rust/pspp/src/dictionary.rs
+++ b/rust/pspp/src/dictionary.rs
@@ -159,8 +159,17 @@ impl Ord for Value {
     }
 }
 
+impl Hash for Value {
+    fn hash<H: std::hash::Hasher>(&self, state: &mut H) {
+        match self {
+            Value::Number(number) => number.map(|x| OrderedFloat(x)).hash(state),
+            Value::String(string) => string.hash(state),
+        }
+    }
+}
+
 impl Value {
-    fn sysmis() -> Self {
+    pub fn sysmis() -> Self {
         Self::Number(None)
     }
 }
diff --git a/rust/pspp/src/identifier.rs b/rust/pspp/src/identifier.rs
index 2d5c0317ec..c6909fd58a 100644
--- a/rust/pspp/src/identifier.rs
+++ b/rust/pspp/src/identifier.rs
@@ -3,7 +3,7 @@ use std::{
     cmp::Ordering,
     fmt::{Debug, Display, Formatter, Result as FmtResult},
     hash::{Hash, Hasher},
-    ops::Deref,
+    ops::{Deref, DerefMut},
 };
 
 use encoding_rs::{EncoderResult, Encoding, UTF_8};
@@ -392,3 +392,12 @@ where
         &self.0
     }
 }
+
+impl<T> DerefMut for ByIdentifier<T>
+where
+    T: HasIdentifier + Clone,
+{
+    fn deref_mut(&mut self) -> &mut Self::Target {
+        &mut self.0
+    }
+}
diff --git a/rust/pspp/src/main.rs b/rust/pspp/src/main.rs
index 5fb57135f7..62ab24337b 100644
--- a/rust/pspp/src/main.rs
+++ b/rust/pspp/src/main.rs
@@ -153,8 +153,8 @@ fn dissect(
             }
             let headers = Headers::new(decoded_records, &|e| eprintln!("{e}"))?;
             let (dictionary, metadata) = decode(headers, encoding, |e| eprintln!("{e}"))?;
-            println!("{dictionary:?}");
-            println!("{metadata:?}");
+            println!("{dictionary:#?}");
+            println!("{metadata:#?}");
         }
     }
 
diff --git a/rust/pspp/src/raw.rs b/rust/pspp/src/raw.rs
index c9b04773ff..5b2a3eb55f 100644
--- a/rust/pspp/src/raw.rs
+++ b/rust/pspp/src/raw.rs
@@ -1553,9 +1553,10 @@ impl ValueLabelRecord<RawStr<8>, RawString> {
         let index_offset = r.stream_position()?;
         let mut dict_indexes = Vec::with_capacity(n as usize);
         let mut invalid_indexes = Vec::new();
+        let valid_range = 1..=var_types.len();
         for _ in 0..n {
             let index: u32 = endian.parse(read_bytes(r)?);
-            if index == 0 || index as usize > var_types.len() {
+            if valid_range.contains(&(index as usize)) {
                 dict_indexes.push(index);
             } else {
                 invalid_indexes.push(index);