add formula reading for xlsb cells_reader

tafia · Nov 1, 2023 · 0c87903 · 0c87903
1 parent 63a5f9e
commit 0c87903
Show file tree

Hide file tree

Showing 8 changed files with 110 additions and 258 deletions.
diff --git a/src/auto.rs b/src/auto.rs
@@ -115,12 +115,12 @@ where
     }
 
     /// 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/lib.rs b/src/lib.rs
@@ -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(),
@@ -230,7 +229,7 @@ where
     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
     ///

diff --git a/src/ods.rs b/src/ods.rs
@@ -185,8 +185,11 @@ where
     }
 
     /// Read worksheet data in corresponding worksheet path
-    fn worksheet_formula(&mut self, name: &str) -> Option<Result<Range<String>, OdsError>> {
-        self.sheets.get(name).map(|r| Ok(r.1.to_owned()))
+    fn worksheet_formula(&mut self, name: &str) -> Result<Range<String>, OdsError> {
+        self.sheets
+            .get(name)
+            .ok_or_else(|| OdsError::WorksheetNotFound(name.into()))
+            .map(|r| r.1.to_owned())
     }
 
     #[cfg(feature = "picture")]

diff --git a/src/xls.rs b/src/xls.rs
@@ -236,8 +236,11 @@ impl<RS: Read + Seek> Reader<RS> for Xls<RS> {
             .collect()
     }
 
-    fn worksheet_formula(&mut self, name: &str) -> Option<Result<Range<String>, XlsError>> {
-        self.sheets.get(name).map(|r| Ok(r.1.clone()))
+    fn worksheet_formula(&mut self, name: &str) -> Result<Range<String>, XlsError> {
+        self.sheets
+            .get(name)
+            .ok_or_else(|| XlsError::WorksheetNotFound(name.into()))
+            .map(|r| r.1.clone())
     }
 
     #[cfg(feature = "picture")]

diff --git a/src/xlsb/cells_reader.rs b/src/xlsb/cells_reader.rs
@@ -5,13 +5,15 @@ use crate::{
     Cell, CellErrorType, Dimensions, XlsbError,
 };
 
-use super::{cell_format, wide_str, RecordIter};
+use super::{cell_format, parse_formula, wide_str, RecordIter};
 
 /// A cells reader for xlsb files
 pub struct XlsbCellsReader<'a> {
     iter: RecordIter<'a>,
     formats: &'a [CellFormat],
     strings: &'a [String],
+    extern_sheets: &'a [String],
+    metadata_names: &'a [(String, String)],
     typ: u16,
     row: u32,
     is_1904: bool,
@@ -24,6 +26,8 @@ impl<'a> XlsbCellsReader<'a> {
         mut iter: RecordIter<'a>,
         formats: &'a [CellFormat],
         strings: &'a [String],
+        extern_sheets: &'a [String],
+        metadata_names: &'a [(String, String)],
         is_1904: bool,
     ) -> Result<Self, XlsbError> {
         let mut buf = Vec::with_capacity(1024);
@@ -55,6 +59,8 @@ impl<'a> XlsbCellsReader<'a> {
             formats,
             is_1904,
             strings,
+            extern_sheets,
+            metadata_names,
             dimensions,
             typ: 0,
             row: 0,
@@ -144,6 +150,52 @@ impl<'a> XlsbCellsReader<'a> {
         let col = read_u32(&self.buf);
         Ok(Some(Cell::new((self.row, col), value)))
     }
+
+    pub fn next_formula(&mut self) -> Result<Option<Cell<String>>, XlsbError> {
+        let value = loop {
+            self.typ = self.iter.read_type()?;
+            let _ = self.iter.fill_buffer(&mut self.buf)?;
+
+            let value = match self.typ {
+                // 0x0001 => continue, // DataType::Empty, // BrtCellBlank
+                0x0008 => {
+                    // BrtFmlaString
+                    let cch = read_u32(&self.buf[8..]) as usize;
+                    let formula = &self.buf[14 + cch * 2..];
+                    let cce = read_u32(formula) as usize;
+                    let rgce = &formula[4..4 + cce];
+                    parse_formula(rgce, &self.extern_sheets, &self.metadata_names)?
+                }
+                0x0009 => {
+                    // BrtFmlaNum
+                    let formula = &self.buf[18..];
+                    let cce = read_u32(formula) as usize;
+                    let rgce = &formula[4..4 + cce];
+                    parse_formula(rgce, &self.extern_sheets, &self.metadata_names)?
+                }
+                0x000A | 0x000B => {
+                    // BrtFmlaBool | BrtFmlaError
+                    let formula = &self.buf[11..];
+                    let cce = read_u32(formula) as usize;
+                    let rgce = &formula[4..4 + cce];
+                    parse_formula(rgce, &self.extern_sheets, &self.metadata_names)?
+                }
+                0x0000 => {
+                    // BrtRowHdr
+                    self.row = read_u32(&self.buf);
+                    if self.row > 0x0010_0000 {
+                        return Ok(None); // invalid row
+                    }
+                    continue;
+                }
+                0x0092 => return Ok(None), // BrtEndSheetData
+                _ => continue, // anything else, ignore and try next, without changing idx
+            };
+            break value;
+        };
+        let col = read_u32(&self.buf);
+        Ok(Some(Cell::new((self.row, col), value)))
+    }
 }
 
 fn parse_dimensions(buf: &[u8]) -> Dimensions {