Hi, I wonder if anyone can help (and please forgive that I don't really understand the things I'm talking about here)
The below is using Stata/IC 14.2 on Windows 7. I cannot see that it has been discussed before on Statalist (though this thread is related)
Take the following code:
A screenshot of test.pdf is shown below. The unicode characters have all failed to render, despite appearing in the Stata graph. Less obvious is that the first {&uarr} is not even replaced by a rectangle, it's just completely absent (which particularly concerns me).

In help whatsnew 13to14, we are told that graphs containing unicode should be exported to PDF, since PS and EPS do not support unicode. But when I export the above to eps, it at least shows me the {&uarr}.
A few questions:
Tim
The below is using Stata/IC 14.2 on Windows 7. I cannot see that it has been discussed before on Statalist (though this thread is related)
Take the following code:
Code:
graph set window fontface "Lucida Sans" tw function 0, tit("←↑→↓↔↕↖↗↘↙↨" "{&uarr}{&uArr}{&larr}{&lArr}{&rarr}{&rArr}{&darr}{&dArr}") translator set Graph2pdf fontface "Lucida Sans" graph export test.pdf, as(pdf)
In help whatsnew 13to14, we are told that graphs containing unicode should be exported to PDF, since PS and EPS do not support unicode. But when I export the above to eps, it at least shows me the {&uarr}.
A few questions:
- How is Stata replacing Lucida Sans in the graph window? (The Lucida Sans does not contain these glyphs)
- Can this replacement be mimicked when exporting to PDF? (I don't think anyone would complain about a change of font to facilitate seeing the text!
- Are there any particularly safe fonts for exporting to PDF?
- Is there any way to tell if there is a problem?
- Will all this be fixed in Stata 15? When will that be?!
Tim
Comment