Merge pull request #370 from tafia/DataTypeRef

support DataTypeRef for shared strings and worksheet CellsReader
tafia · Dec 12, 2023 · b72eb4f · b72eb4f
2 parents 7342e2d + aa19868
commit b72eb4f
Show file tree

Hide file tree

Showing 15 changed files with 1,020 additions and 779 deletions.
diff --git a/benches/basic.rs b/benches/basic.rs
@@ -17,7 +17,6 @@ fn count<R: Reader<BufReader<File>>>(path: &str) -> usize {
         count += excel
             .worksheet_range(&s)
             .unwrap()
-            .unwrap()
             .rows()
             .flat_map(|r| r.iter())
             .count();
@@ -44,3 +43,41 @@ fn bench_xlsb(b: &mut Bencher) {
 fn bench_ods(b: &mut Bencher) {
     b.iter(|| count::<Ods<_>>("tests/issues.ods"));
 }
+
+#[bench]
+fn bench_xlsx_cells_reader(b: &mut Bencher) {
+    fn count<R: Reader<BufReader<File>>>(path: &str) -> usize {
+        let path = format!("{}/{}", env!("CARGO_MANIFEST_DIR"), path);
+        let mut excel: Xlsx<_> = open_workbook(&path).expect("cannot open excel file");
+
+        let sheets = excel.sheet_names().to_owned();
+        let mut count = 0;
+        for s in sheets {
+            let mut cells_reader = excel.worksheet_cells_reader(&s).unwrap();
+            while let Some(_) = cells_reader.next_cell().unwrap() {
+                count += 1;
+            }
+        }
+        count
+    }
+    b.iter(|| count::<Xlsx<_>>("tests/issues.xlsx"));
+}
+
+#[bench]
+fn bench_xlsb_cells_reader(b: &mut Bencher) {
+    fn count<R: Reader<BufReader<File>>>(path: &str) -> usize {
+        let path = format!("{}/{}", env!("CARGO_MANIFEST_DIR"), path);
+        let mut excel: Xlsb<_> = open_workbook(&path).expect("cannot open excel file");
+
+        let sheets = excel.sheet_names().to_owned();
+        let mut count = 0;
+        for s in sheets {
+            let mut cells_reader = excel.worksheet_cells_reader(&s).unwrap();
+            while let Some(_) = cells_reader.next_cell().unwrap() {
+                count += 1;
+            }
+        }
+        count
+    }
+    b.iter(|| count::<Xlsx<_>>("tests/issues.xlsb"));
+}
diff --git a/examples/excel_to_csv.rs b/examples/excel_to_csv.rs
@@ -24,7 +24,7 @@ fn main() {
     let dest = sce.with_extension("csv");
     let mut dest = BufWriter::new(File::create(dest).unwrap());
     let mut xl = open_workbook_auto(&sce).unwrap();
-    let range = xl.worksheet_range(&sheet).unwrap().unwrap();
+    let range = xl.worksheet_range(&sheet).unwrap();
 
     write_range(&mut dest, &range).unwrap();
 }

diff --git a/examples/search_errors.rs b/examples/search_errors.rs
@@ -74,10 +74,7 @@ fn run(f: GlobResult) -> Result<(PathBuf, Option<usize>, usize), FileStatus> {
     let sheets = xl.sheet_names().to_owned();
 
     for s in sheets {
-        let range = xl
-            .worksheet_range(&s)
-            .unwrap()
-            .map_err(FileStatus::RangeError)?;
+        let range = xl.worksheet_range(&s).map_err(FileStatus::RangeError)?;
         cell_errors += range
             .rows()
             .flat_map(|r| {

diff --git a/src/auto.rs b/src/auto.rs
@@ -105,22 +105,22 @@ where
     }
 
     /// Read worksheet data in corresponding worksheet path
-    fn worksheet_range(&mut self, name: &str) -> Option<Result<Range<DataType>, Self::Error>> {
+    fn worksheet_range(&mut self, name: &str) -> Result<Range<DataType>, Self::Error> {
         match *self {
-            Sheets::Xls(ref mut e) => e.worksheet_range(name).map(|r| r.map_err(Error::Xls)),
-            Sheets::Xlsx(ref mut e) => e.worksheet_range(name).map(|r| r.map_err(Error::Xlsx)),
-            Sheets::Xlsb(ref mut e) => e.worksheet_range(name).map(|r| r.map_err(Error::Xlsb)),
-            Sheets::Ods(ref mut e) => e.worksheet_range(name).map(|r| r.map_err(Error::Ods)),
+            Sheets::Xls(ref mut e) => e.worksheet_range(name).map_err(Error::Xls),
+            Sheets::Xlsx(ref mut e) => e.worksheet_range(name).map_err(Error::Xlsx),
+            Sheets::Xlsb(ref mut e) => e.worksheet_range(name).map_err(Error::Xlsb),
+            Sheets::Ods(ref mut e) => e.worksheet_range(name).map_err(Error::Ods),
         }
     }
 
     /// Read worksheet formula in corresponding worksheet path
-    fn worksheet_formula(&mut self, name: &str) -> Option<Result<Range<String>, Self::Error>> {
+    fn worksheet_formula(&mut self, name: &str) -> Result<Range<String>, Self::Error> {
         match *self {
-            Sheets::Xls(ref mut e) => e.worksheet_formula(name).map(|r| r.map_err(Error::Xls)),
-            Sheets::Xlsx(ref mut e) => e.worksheet_formula(name).map(|r| r.map_err(Error::Xlsx)),
-            Sheets::Xlsb(ref mut e) => e.worksheet_formula(name).map(|r| r.map_err(Error::Xlsb)),
-            Sheets::Ods(ref mut e) => e.worksheet_formula(name).map(|r| r.map_err(Error::Ods)),
+            Sheets::Xls(ref mut e) => e.worksheet_formula(name).map_err(Error::Xls),
+            Sheets::Xlsx(ref mut e) => e.worksheet_formula(name).map_err(Error::Xlsx),
+            Sheets::Xlsb(ref mut e) => e.worksheet_formula(name).map_err(Error::Xlsb),
+            Sheets::Ods(ref mut e) => e.worksheet_formula(name).map_err(Error::Ods),
         }
     }
 

diff --git a/src/datatype.rs b/src/datatype.rs
@@ -360,6 +360,53 @@ where
     }
 }
 
+/// An enum to represent all different data types that can appear as
+/// a value in a worksheet cell
+#[derive(Debug, Clone, PartialEq, Default)]
+pub enum DataTypeRef<'a> {
+    /// Signed integer
+    Int(i64),
+    /// Float
+    Float(f64),
+    /// String
+    String(String),
+    /// Shared String
+    SharedString(&'a str),
+    /// Boolean
+    Bool(bool),
+    /// Date or Time
+    DateTime(f64),
+    /// Duration
+    Duration(f64),
+    /// Date, Time or DateTime in ISO 8601
+    DateTimeIso(String),
+    /// Duration in ISO 8601
+    DurationIso(String),
+    /// Error
+    Error(CellErrorType),
+    /// Empty cell
+    #[default]
+    Empty,
+}
+
+impl<'a> From<DataTypeRef<'a>> for DataType {
+    fn from(value: DataTypeRef<'a>) -> Self {
+        match value {
+            DataTypeRef::Int(v) => DataType::Int(v),
+            DataTypeRef::Float(v) => DataType::Float(v),
+            DataTypeRef::String(v) => DataType::String(v),
+            DataTypeRef::SharedString(v) => DataType::String(v.into()),
+            DataTypeRef::Bool(v) => DataType::Bool(v),
+            DataTypeRef::DateTime(v) => DataType::DateTime(v),
+            DataTypeRef::Duration(v) => DataType::Duration(v),
+            DataTypeRef::DateTimeIso(v) => DataType::DateTimeIso(v),
+            DataTypeRef::DurationIso(v) => DataType::DurationIso(v),
+            DataTypeRef::Error(v) => DataType::Error(v),
+            DataTypeRef::Empty => DataType::Empty,
+        }
+    }
+}
+
 #[cfg(all(test, feature = "dates"))]
 mod date_tests {
     use super::*;

diff --git a/src/de.rs b/src/de.rs
@@ -110,8 +110,7 @@ impl RangeDeserializerBuilder<'static, &'static str> {
     /// fn main() -> Result<(), Error> {
     ///     let path = format!("{}/tests/temperature.xlsx", env!("CARGO_MANIFEST_DIR"));
     ///     let mut workbook: Xlsx<_> = open_workbook(path)?;
-    ///     let range = workbook.worksheet_range("Sheet1")
-    ///         .ok_or(Error::Msg("Cannot find 'Sheet1'"))??;
+    ///     let range = workbook.worksheet_range("Sheet1")?;
     ///
     ///     let mut iter = RangeDeserializerBuilder::new()
     ///         .has_headers(false)
@@ -154,8 +153,7 @@ impl<'h, H: AsRef<str> + Clone + 'h> RangeDeserializerBuilder<'h, H> {
     /// fn main() -> Result<(), Error> {
     ///     let path = format!("{}/tests/temperature.xlsx", env!("CARGO_MANIFEST_DIR"));
     ///     let mut workbook: Xlsx<_> = open_workbook(path)?;
-    ///     let range = workbook.worksheet_range("Sheet1")
-    ///         .ok_or(Error::Msg("Cannot find 'Sheet1'"))??;
+    ///     let range = workbook.worksheet_range("Sheet1")?;
     ///     let mut iter = RangeDeserializerBuilder::with_headers(&["value", "label"]).from_range(&range)?;
     ///
     ///     if let Some(result) = iter.next() {
@@ -184,8 +182,7 @@ impl<'h, H: AsRef<str> + Clone + 'h> RangeDeserializerBuilder<'h, H> {
     /// fn main() -> Result<(), Error> {
     ///     let path = format!("{}/tests/temperature.xlsx", env!("CARGO_MANIFEST_DIR"));
     ///     let mut workbook: Xlsx<_> = open_workbook(path)?;
-    ///     let range = workbook.worksheet_range("Sheet1")
-    ///         .ok_or(Error::Msg("Cannot find 'Sheet1'"))??;
+    ///     let range = workbook.worksheet_range("Sheet1")?;
     ///     let mut iter = RangeDeserializerBuilder::new().from_range(&range)?;
     ///
     ///     if let Some(result) = iter.next() {
@@ -220,8 +217,7 @@ impl<'h, H: AsRef<str> + Clone + 'h> RangeDeserializerBuilder<'h, H> {
 /// fn main() -> Result<(), Error> {
 ///     let path = format!("{}/tests/temperature.xlsx", env!("CARGO_MANIFEST_DIR"));
 ///     let mut workbook: Xlsx<_> = open_workbook(path)?;
-///     let range = workbook.worksheet_range("Sheet1")
-///         .ok_or(Error::Msg("Cannot find 'Sheet1'"))??;
+///     let range = workbook.worksheet_range("Sheet1")?;
 ///
 ///     let mut iter = RangeDeserializerBuilder::new().from_range(&range)?;
 ///

diff --git a/src/formats.rs b/src/formats.rs
@@ -1,4 +1,4 @@
-use crate::DataType;
+use crate::{datatype::DataTypeRef, DataType};
 
 /// https://learn.microsoft.com/en-us/office/troubleshoot/excel/1900-and-1904-date-system
 static EXCEL_1900_1904_DIFF: i64 = 1462;
@@ -104,18 +104,28 @@ pub fn format_excel_i64(value: i64, format: Option<&CellFormat>, is_1904: bool)
 }
 
 // convert f64 to date, if format == Date
-pub fn format_excel_f64(value: f64, format: Option<&CellFormat>, is_1904: bool) -> DataType {
+#[inline]
+pub fn format_excel_f64_ref<'a>(
+    value: f64,
+    format: Option<&CellFormat>,
+    is_1904: bool,
+) -> DataTypeRef<'static> {
     match format {
-        Some(CellFormat::DateTime) => DataType::DateTime(if is_1904 {
+        Some(CellFormat::DateTime) => DataTypeRef::DateTime(if is_1904 {
             value + EXCEL_1900_1904_DIFF as f64
         } else {
             value
         }),
-        Some(CellFormat::TimeDelta) => DataType::Duration(value),
-        _ => DataType::Float(value),
+        Some(CellFormat::TimeDelta) => DataTypeRef::Duration(value),
+        _ => DataTypeRef::Float(value),
     }
 }
 
+// convert f64 to date, if format == Date
+pub fn format_excel_f64(value: f64, format: Option<&CellFormat>, is_1904: bool) -> DataType {
+    format_excel_f64_ref(value, format, is_1904).into()
+}
+
 /// Ported from openpyxl, MIT License
 /// https://foss.heptapod.net/openpyxl/openpyxl/-/blob/a5e197c530aaa49814fd1d993dd776edcec35105/openpyxl/styles/tests/test_number_style.py
 #[test]

diff --git a/src/lib.rs b/src/lib.rs
@@ -15,7 +15,7 @@
 //! let mut workbook: Xlsx<_> = open_workbook(path).expect("Cannot open file");
 //!
 //! // Read whole worksheet data and provide some statistics
-//! if let Some(Ok(range)) = workbook.worksheet_range("Sheet1") {
+//! if let Ok(range) = workbook.worksheet_range("Sheet1") {
 //!     let total_cells = range.get_size().0 * range.get_size().1;
 //!     let non_empty_cells: usize = range.used_cells().count();
 //!     println!("Found {} cells in 'Sheet1', including {} non empty cells",
@@ -49,7 +49,6 @@
 //!     println!("found {} formula in '{}'",
 //!              workbook
 //!                 .worksheet_formula(&s)
-//!                 .expect("sheet not found")
 //!                 .expect("error while getting formula")
 //!                 .rows().flat_map(|r| r.iter().filter(|f| !f.is_empty()))
 //!                 .count(),
@@ -74,6 +73,7 @@ mod de;
 mod errors;
 pub mod vba;
 
+use datatype::DataTypeRef;
 use serde::de::DeserializeOwned;
 use std::borrow::Cow;
 use std::cmp::{max, min};
@@ -132,6 +132,18 @@ impl fmt::Display for CellErrorType {
     }
 }
 
+#[derive(Debug, PartialEq, Default, Clone, Copy)]
+pub(crate) struct Dimensions {
+    pub start: (u32, u32),
+    pub end: (u32, u32),
+}
+
+impl Dimensions {
+    pub fn len(&self) -> u64 {
+        (self.end.0 - self.start.0 + 1) as u64 * (self.end.1 - self.start.1 + 1) as u64
+    }
+}
+
 /// Common file metadata
 ///
 /// Depending on file type, some extra information may be stored
@@ -203,18 +215,21 @@ where
 
     /// Creates a new instance.
     fn new(reader: RS) -> Result<Self, Self::Error>;
+
     /// Gets `VbaProject`
     fn vba_project(&mut self) -> Option<Result<Cow<'_, VbaProject>, Self::Error>>;
+
     /// Initialize
     fn metadata(&self) -> &Metadata;
+
     /// Read worksheet data in corresponding worksheet path
-    fn worksheet_range(&mut self, name: &str) -> Option<Result<Range<DataType>, Self::Error>>;
+    fn worksheet_range(&mut self, name: &str) -> Result<Range<DataType>, Self::Error>;
 
     /// Fetch all worksheet data & paths
     fn worksheets(&mut self) -> Vec<(String, Range<DataType>)>;
 
     /// Read worksheet formula in corresponding worksheet path
-    fn worksheet_formula(&mut self, _: &str) -> Option<Result<Range<String>, Self::Error>>;
+    fn worksheet_formula(&mut self, _: &str) -> Result<Range<String>, Self::Error>;
 
     /// Get all sheet names of this workbook, in workbook order
     ///
@@ -248,7 +263,7 @@ where
     /// sheet_name, then the corresponding worksheet.
     fn worksheet_range_at(&mut self, n: usize) -> Option<Result<Range<DataType>, Self::Error>> {
         let name = self.sheet_names().get(n)?.to_string();
-        self.worksheet_range(&name)
+        Some(self.worksheet_range(&name))
     }
 
     /// Get all pictures, tuple as (ext: String, data: Vec<u8>)
@@ -279,6 +294,7 @@ where
 pub trait CellType: Default + Clone + PartialEq {}
 
 impl CellType for DataType {}
+impl<'a> CellType for DataTypeRef<'a> {}
 impl CellType for String {}
 impl CellType for usize {} // for tests
 
@@ -616,8 +632,7 @@ impl<T: CellType> Range<T> {
     /// fn main() -> Result<(), Error> {
     ///     let path = format!("{}/tests/temperature.xlsx", env!("CARGO_MANIFEST_DIR"));
     ///     let mut workbook: Xlsx<_> = open_workbook(path)?;
-    ///     let mut sheet = workbook.worksheet_range("Sheet1")
-    ///         .ok_or(Error::Msg("Cannot find 'Sheet1'"))??;
+    ///     let mut sheet = workbook.worksheet_range("Sheet1")?;
     ///     let mut iter = sheet.deserialize()?;
     ///
     ///     if let Some(result) = iter.next() {