[Octave-bug-tracker] [bug #51512] [octave forge] (io) Missing or wrong types when using xlsread with OCT interface

[Markus Mützel]

Follow-up Comment #19, bug #51512 (project octave):

The last version seems to run at more or less the same speed as the version
before.
Your file (file #41434):

>> profile clear
>> profile on
>> tic; [~, ~, raw] = xlsread ("Excel2013_1001x999.xlsx", 1, "", "oct"); toc
combined regexps
Elapsed time is 95.3975 seconds.
>> profile off
>> profshow
   #         Function Attr     Time (s)   Time (%)        Calls
---------------------------------------------------------------
  54           regexp            62.753      66.04         1123
  62       str2double             8.586       9.04            8
  60              cat             4.837       5.09         1129
  57          cellfun             3.533       3.72        10065
  83          col2num             2.999       3.16      1999998
  70 __OCT_xlsx2oct__             2.996       3.15            1
  79          strncmp             1.782       1.88            5
  67          xls2oct             1.431       1.51            1
  78        regexprep             1.206       1.27            6
  47           system             1.201       1.26            1
  87           strrep             1.032       1.09            5
  72         num2cell             0.790       0.83            2
  55         cell2mat             0.787       0.83         2234
   5           ischar             0.302       0.32      1000038
  66        postfix '             0.114       0.12            6
  27           strcmp             0.107       0.11           25
  77        iscellstr             0.077       0.08            6
  88        parsecell             0.073       0.08            1
  82            clear             0.071       0.07            1
  19             cell             0.066       0.07            6


My file (file #41352 + changes from comment #16):

>> profile clear
>> profile on
>> tic; [~, ~, raw] = xlsread ("Excel2013_1001x999.xlsx", 1, "", "oct"); toc
Elapsed time is 94.7151 seconds.
>> profile off
>> profshow
   #         Function Attr     Time (s)   Time (%)        Calls
---------------------------------------------------------------
  54           regexp            64.371      68.22         1128
  62       str2double             9.943      10.54            8
  60              cat             4.721       5.00         1132
  57          cellfun             3.547       3.76        10083
  77          col2num             3.005       3.19      1999998
  70 __OCT_xlsx2oct__             2.770       2.94            1
  47           system             1.168       1.24            1
  67          xls2oct             1.109       1.17            1
  81           strrep             0.940       1.00            5
  55         cell2mat             0.781       0.83         2239
  71         num2cell             0.772       0.82            5
  74            clear             0.517       0.55            6
   5           ischar             0.312       0.33      1000022
  66        postfix '             0.081       0.09            9
  82        parsecell             0.073       0.08            1


Immediately followed by "COM":

>> tic; [~, ~, raw] = xlsread ("Excel2013_1001x999.xlsx", 1, "", "com"); toc

Checking requested interface(s):
COM*;
Elapsed time is 4.22336 seconds.


I could not find any reference to a date/time type in the specifications. The
only available types I found are:

<simpleType name="ST_CellType">
 <restriction base="xsd:string">
  <enumeration value="b"/>
  <enumeration value="n"/>
  <enumeration value="e"/>
  <enumeration value="s"/>
  <enumeration value="str"/>
  <enumeration value="inlineStr"/>
 </restriction>
</simpleType>


Dates are stored as doubles in my tests. But I suppose it doesn't hurt to
leave the logic in if you have seen that type in some file.

In the attached file, I tried to replace any "." match with a negative match
to avoid the trouble with PCRE's MATCH_LIMIT. Additionally, I changed slightly
how the matched results are checked for consistency such that multiples of 3
errors can also be detected.
The file still performs approximately identically:

>> profile clear
>> profile on
>> tic; [~, ~, raw] = xlsread ("Excel2013_1001x999.xlsx", 1, "", "oct"); toc
combined regexps
Elapsed time is 93.4514 seconds.
>> profile off
>> profshow
   #         Function Attr     Time (s)   Time (%)        Calls
---------------------------------------------------------------
  54           regexp            62.862      67.55         1123
  62       str2double             8.520       9.15            8
  60              cat             4.825       5.18         1129
  82          col2num             3.031       3.26      1999998
  57          cellfun             2.868       3.08        10060
  70 __OCT_xlsx2oct__             2.762       2.97            1
  77          strncmp             1.742       1.87            5
  67          xls2oct             1.415       1.52            1
  76        regexprep             1.192       1.28            6
  47           system             1.159       1.25            1
  86           strrep             1.007       1.08            5
  55         cell2mat             0.689       0.74         2233
   5           ischar             0.330       0.35      1000038
  27           strcmp             0.108       0.12           25
  75        iscellstr             0.078       0.08            6
  81            clear             0.072       0.08            1
  87        parsecell             0.071       0.08            1
  19             cell             0.066       0.07            6
  49            fread             0.063       0.07            4
   2          xlsread             0.051       0.05            1



(file #41505)
    _______________________________________________________

Additional Item Attachment:

File name: __OCT_xlsx2oct__.m             Size:9 KB


    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?51512>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/