just one classifier choice for each single symbol is created

[namv...@gmail.com]

Hi,
I'm Trying to find classifier choices for a single symbol by using api in wiki page(https://github.com/tesseract-ocr/tesseract/wiki/APIExample)
the results for each character in English is like this:
symbol : c  conf:89.6814
c conf:89.6814
e conf:83.7891
C conf:80.4431
¢ conf:77.6685

When l used the traineddata which I created by myself, the result was the same.
but when I use ara.traineddata for Arabic language  (downloaded from https://github.com/tesseract-ocr/tessdata)
It shows only one choice for each character. how can I fix this to get multiple results for each character in Arabic too?