| T O P I C R E V I E W |
| cvhorie |
Posted - 09/12/2009 : 11:00:43 AM
Origin Ver. and Service Release 8 sr6
Operating System: XP
25 years ago, I plotted this graph (and 30 similar) using Fortran and punched cards. I am now replotting them with new data. Origin is much easier, using the Dataset option to filter the data to be displayed and how.
A number of questions remain.
1. Grid lines are always confusing on ternary graphs, so I stop them before they get to the opposite axis. How do I control the length/position of the grid lines in Origin?
2. I have set up the legend as below using the enhanced legend, with the symbol edge thickness in tools>options>graph set to 25%. But the open diamond (5) and cross (7) symbols are displayed with very thin edges, unlike the same symbols that are plotted in the graph.
\l(s(0,2,0,1,1,6,0,1,0,50,25))soluble
\l(s(0,5,3,1,1,6,0,1,0,50,25))borderline
\l(s(0,7,5,1,1,5,0,1,0,50,-1))insoluble
3. I want to use the dataset (not dataplot) column heading (long name) as the title for the graph. But I cannot find a list of the appropriate variables that can be used in a text box to refer to this. Where are the various reference lists? They should be in an appendix in the manual.
4. In the Dataset, the categorical information (e.g. a code for soluble) is related to a particular sample which is identified by a reference number by row in the worksheet. How can these references be extracted and listed out, as in the example?
5. Is there a structured learning course for the Labtalk programming language?
Thank you for your help.
Velson Horie |
| 6 L A T E S T R E P L I E S (Newest First) |
| cvhorie |
Posted - 09/22/2009 : 02:54:22 AM
Thank you Amanda for identifying the problems with labelling. Much easier to cope when these are clear. The revised syntax works fine.
I should be grateful if you could send me the dataset and plot for the ternary you created. I am still struggling with the 400 page LT manual and need pointers about how to configure and use variables.
Web interface:
The last couple of posts have disappeared from this thread. Is that normal?
Whenever I start to search the forum each time, I get an error message from the server. When I then do an identical re-search, the proper response appears.
Velson Horie |
| AmandaLu |
Posted - 09/21/2009 : 07:26:13 AM
Hi,
Origin has no grouping feature for plotting right now. So you did need to control the legend display by LT script. We got many similar requests like this, so in our future version, Origin 8.1, we will introduce unstack columns tool to reorganize data automatically.
About the legend display problem, I am sorry that there is problem in our documents.
Here are the correct syntaxes:
\L(O SymEdgeColor,Sym,Fill,Size,ColorLn,LineStyle,Gap,LnWidth,SymFillColor, EdgeThickness)
and
\L(S Category, arg2, arg3, arg4, arg5, arg6, LineStyle, LineColor, LineWidth, Gap, EdgeThickness)
|
| cvhorie |
Posted - 09/18/2009 : 1:19:31 PM
This works, but it is a bit flakey because the box does not like being opened again.
Enhanced legend has 11 arguments
\L(S(Category, arg2, arg3, arg4, arg5, arg6, LineStyle, LineColor, LineWidth, Gap, EdgeThickness))
Advanced legend has 10 arguments
\L(O(SymEdgeColor,Sym,Fill,Size,ColorLn,LineStyle,Gap,LnWidth,SymFillColor, EdgeThickness))
The method you suggested seems to depend on the separate columns being designated and chosen as data sets. But in the data list, there is only one entry, which was created when I chose ternary as the graph type. A similar requirement appears to be needed to locate the individual row/column designations for identifying then categorising the data to output the numbers in the text.
Would you be able to send me the data/graph you constructed to check whether we have our data in comparable format?
Velson Horie |
| cvhorie |
Posted - 09/18/2009 : 10:50:25 AM
Legend difficulty
On the help page - Manual Control of the Graph Legend
the syntax is
\L(O(SymEdgeColor,Sym,Fill,Size,ColorLn,LineStyle,Gap,LnWidth,SymFillColor, EdgeThickness))
10 arguments
In the manual and on the help page Enhanced legend
\L(S(Category, arg2, arg3, arg4, arg5, arg6, LineStyle, LineColor, LineWidth, Gap, EdgeThickness))
has 11 arguments, so can you spot where I might have gone wrong?
Interestingly
\l(s(0,2,0,1,1,6,0,1,0,50,25))soluble
\l(o(1,5,3,6,1,0,50,0,1,50
borderline
\l(s(0,7,5,1,1,5,0,1,0,50,50
))insoluble
produces a thick walled diamond symbol and cross as required, but with the text on the line below. But if ))borderline" is placed at the end of the second line (after the command), the diamond reverts to thin walled.
Your proposal for using
\l(1) %(1)
\l(2) %(2)
\l(3) %(3)
seems to rely on the data being held in three separate data groups which are referred to with \L(1) etc. However when I used the ternary tool to create the graph, all three sets of data are put in one group - there is only one entry at the bottom of the list of data.
Presumably the same requirement is needed to identify the data required to sort the data into categories for outputting in relationship to the symbols.
Would it be possible for you to send the graph you created to check if we are working with data in comparable formats?
Thank you
Velson Horie |
| cvhorie |
Posted - 09/18/2009 : 08:51:38 AM
Amanda
I am really grateful for the help.
I had already tried to adjust the ticks as you suggest, but yours look better. Thank you for taking up the idea of a shortened gridline.
I shall have to be more careful in my programming skills (and assume that I am more likely to be wrong) and get the no of arguments right.
Your script will save me a lot struggling. My task for the weekend is to read, digest and apply the LabTalk language to this problem. You have given me a head start in the learning curve, reverse engineering the commands and so concentrate on what is relevant.
Thank you
Velson Horie |
| AmandaLu |
Posted - 09/14/2009 : 07:42:39 AM
Hi,
1. I am afraid that Origin does not support it yet. But thank you for your suggestion, it is a good point. We have submitted a tracker, #14322, for it and will have it implement in our future version.
By the way, I made a ternary in Origin showing in the picture below. You may notice that our major ticks align in the same direction as grid lines so one can easily identify a tick and its corresponding grid line without confusing. Do you think it is fine enough?
2. The enhanced legend syntax has only 10 variables but your legend script has 11. Perhaps this is why yours does not work. Anyway, what about using our default legend?
\l(1) %(1)
\l(2) %(2)
\l(3) %(3)
4. This question can be done by a small LT script. Since I do not have your data, I make some data to show you an example. The original data is in the book3. The categorical information is stored in col(J).
I use the following script to add a text label to the graph:
range r1=[book3]"Ternary1"!col(J);%a="";
for (ii=1; ii<=r1.nrows; ii++)
{
%A=%A %(r1[ii]$);//extract the values in col(J) and store them in %A
}
%B=r1[L]$;//store col(J)¡¯s long name in %B
label -s -n myLabel %B %A;//create a text label of col(J)
5.To learn Labtalk Programming, you can refer to our LT Guide in 81. We are still working for it. But it should be read for learning.
http://wiki.originlab.com/~originla/wiki/index.php?title=Category:LT_Guide
Amanda
OriginLab |
|
|
