sheetjs-clone/bits/72_wsxml.js
SheetJS a96d8964e2 version bump 0.5.0: changes to cell interface
Cell object stores the formatted text in the `.w` field
- Raw format and types are preserved in the `.t`, `.v` fields
- Accessors have been updated to use the field

updates for older OOXML:
- ECMA-376 makes reference to old sstItem shared string item
- [MS-XLSX] xmlns for workbook
2014-02-03 19:00:44 -05:00

82 lines
2.7 KiB
JavaScript

/* 18.3 Worksheets */
function parse_worksheet(data) {
if(!data) return data;
/* 18.3.1.99 worksheet CT_Worksheet */
var s = {};
/* 18.3.1.35 dimension CT_SheetDimension ? */
var ref = data.match(/<dimension ref="([^"]*)"\s*\/>/);
if(ref && ref.length == 2 && ref[1].indexOf(":") !== -1) s["!ref"] = ref[1];
var refguess = {s: {r:1000000, c:1000000}, e: {r:0, c:0} };
var q = ["v","f"];
var sidx = 0;
/* 18.3.1.80 sheetData CT_SheetData ? */
if(!data.match(/<sheetData *\/>/))
data.match(/<sheetData>([^\u2603]*)<\/sheetData>/m)[1].split("</row>").forEach(function(x) {
if(x === "" || x.trim() === "") return;
/* 18.3.1.73 row CT_Row */
var row = parsexmltag(x.match(/<row[^>]*>/)[0]);
if(refguess.s.r > row.r - 1) refguess.s.r = row.r - 1;
if(refguess.e.r < row.r - 1) refguess.e.r = row.r - 1;
/* 18.3.1.4 c CT_Cell */
var cells = x.substr(x.indexOf('>')+1).split(/<c /);
cells.forEach(function(c, idx) { if(c === "" || c.trim() === "") return;
var cref = c.match(/r="([^"]*)"/);
c = "<c " + c;
if(cref && cref.length == 2) {
var cref_cell = decode_cell(cref[1]);
idx = cref_cell.c;
}
if(refguess.s.c > idx) refguess.s.c = idx;
if(refguess.e.c < idx) refguess.e.c = idx;
var cell = parsexmltag((c.match(/<c[^>]*>/)||[c])[0]); delete cell[0];
var d = c.substr(c.indexOf('>')+1);
var p = {};
q.forEach(function(f){var x=d.match(matchtag(f));if(x)p[f]=unescapexml(x[1]);});
/* SCHEMA IS ACTUALLY INCORRECT HERE. IF A CELL HAS NO T, EMIT "" */
if(cell.t === undefined && p.v === undefined) { p.t = "str"; p.v = undefined; }
else p.t = (cell.t ? cell.t : "n"); // default is "n" in schema
switch(p.t) {
case 'n': p.v = parseFloat(p.v); break;
case 's': {
sidx = parseInt(p.v, 10);
p.v = strs[sidx].t;
p.r = strs[sidx].r;
} break;
case 'str': if(p.v) p.v = utf8read(p.v); break;
case 'inlineStr':
var is = d.match(/<is>(.*)<\/is>/);
is = is ? parse_si(is[1]) : {t:"",r:""};
p.t = 'str'; p.v = is.t;
break; // inline string
case 'b':
switch(p.v) {
case '0': case 'FALSE': case "false": case false: p.v=false; break;
case '1': case 'TRUE': case "true": case true: p.v=true; break;
default: throw "Unrecognized boolean: " + p.v;
} break;
/* in case of error, stick value in .raw */
case 'e': p.raw = RBErr[p.v]; break;
default: throw "Unrecognized cell type: " + p.t;
}
/* formatting */
var fmtid = 0;
if(cell.s && styles.CellXf) {
var cf = styles.CellXf[cell.s];
if(cf && cf.numFmtId) fmtid = cf.numFmtId;
}
try { p.w = SSF.format(fmtid,p.v,_ssfopts); } catch(e) { }
s[cell.r] = p;
});
});
if(!s["!ref"]) s["!ref"] = encode_range(refguess);
return s;
}