v1.3 fixed repeater bug

R/readODS.R

@@ -6,7 +6,7 @@ library(XML)
 # 
 #TODO: test gnummeric and koffice generated ODS files
 #TODO: check if emtpy rows are the only ones with "number-rows-repeated"...
-# ALSO number-columns-repeated
+# ALSO number-columns-repeated -- FIXED
 #
 # the thing i should have probably read (AKA the ODS standard :P): 
 # http://docs.oasis-open.org/office/v1.2/os/OpenDocument-v1.2-os-part1.html#__RefHeading__1420324_253892949
@@ -49,14 +49,13 @@ read.ods=function(file=NULL, sheet=NULL, formulaAsFormula=F){
   for(sheeti in sheets[names(sheets)=="table"]){
     sheetIndex=sheetIndex+1
     #sheeti=sheets[[3]]
-
+    
     d=list()
     d[1]="" # avoid bug later on if no rows
     # fill it
     rowIndex=0
     for(row in sheeti[names(sheeti)=="table-row"]){
       rowIndex=rowIndex+1
-#       print(rowIndex)
       if(!is.na(xmlAttrs(row)["number-rows-repeated"])){
         # only on empty rows???
         rowIndex=rowIndex+as.integer(xmlAttrs(row)[["number-rows-repeated"]])-1
@@ -69,17 +68,12 @@ read.ods=function(file=NULL, sheet=NULL, formulaAsFormula=F){
         if(is.null(xmlAttrs(cell))){ # silly liblre office has: <table:table-cell/>
           next
         }
-#         print(colIndex)
-        print(paste("row:",rowIndex," col:",colIndex,sep=""))
+#         print(paste("row:",rowIndex," col:",colIndex,sep=""))
         #<table:table-cell table:number-columns-repeated="3"/>
-#         colsRepeated=1
-#        print(names(cell))
-#        print(length((names(cell))))
         if(!is.na(xmlAttrs(cell)["number-columns-repeated"]) && length((names(cell)))==0  ){
 #           print(as.integer(xmlAttrs(cell)[["number-columns-repeated"]]))
           # repeat empty columns
           colIndex=colIndex+as.integer(xmlAttrs(cell)[["number-columns-repeated"]])-1
-#           colsRepeated=as.integer(xmlAttrs(cell)[["number-columns-repeated"]]
           next
         }
         # display the formula instead of its result
@@ -265,6 +259,8 @@ lettersToNumber=function( listOfStrings=NULL){
 #' libre office can do crap like this:
 #' <text:p>></text:p>
 #' which is not valid xml... and the XML package doesn't like that..
+#' OK NM NOW IT DOES! leaving the code here in case of changes...
+#' 
 #' also to not make the code fulgy as all hell, the content.xml file is first scanned for these XML violations, and then fixed!
 #' <text:p>></text:p> --> <text:p>&gt</text:p>
 #' @param file - the xml file to be parsed...

notes.R

@@ -33,6 +33,7 @@ read.ods(file,formulaAsFormula = T)
 read.ods(file ,sheet=1)
 
 
+
 read.ods(file,usePreParser = F)
 
 odsPreParser(file)

tests/testthat/test_readODS.R

@@ -44,7 +44,33 @@ test_that("read.ods", {
                 stringsAsFactors = F)
   expect_equal(read.ods(file, sheet=1),df)
 
+  file=paste("../testdata/layout_test.ods",sep="")
+  sheet1=read.ods(file, sheet=1)
+  expect_equal(sheet1[8,"F"],"empty") # this is a repeated element
 
+  sheet2=read.ods(file, sheet=2)
+  expect_equal(dim(sheet2),c(22,13))
+  expect_true(all(sheet1[21,]==sheet2[22,]))
+
+  file=paste("../testdata/multisheet.ods",sep="")
+  df=data.frame(matrix("",14,7),stringsAsFactors = F)
+  df[1,1]="COOKIES"
+  df[4,2]="1"
+  df[6,3]="2"
+  df[8,3]="3"
+  df[14,4]="3"
+  df[7,5]="3"
+  df[9,5]="1"
+  df[10,7]="1"
+  sheet2=read.ods(file, sheet=2)
+  expect_true(all(sheet2==df))
+
+
+
+  # this test will fail on windows!!!
+#   file=paste("../testdata/test.ods",sep="")
+#   sheet1=read.ods(file, sheet=1)
+#   expect_equal(sheet1[3,"E"],'lalala lalala,,””,,') 
 
 })