Problem Figuring Out Why Percentage for a Particular Group Fails to Equal 1

[MD Mubtasim-Fuad]

So I have this graph created in the bar graph editor, based off election survey data. I restricted all observations in the dependent variable to the three major parties listed, with all else recorded as missing. My three indep. vars here, region, visible minority status and immigration status, are all restricted to drop the observation if it conflicts with at least one or the other variable conditions or is missing (I also selected the exclude missing variables option in the graph editor). I am not sure if STATA's graph editor auto-drops an observation if at least one of the dep./indep. vars are missing. Now, I am confused why for some regions, the respective group (i.e. Canadian-born of European descent) adds up to 1, such as in the Prairies, but for others, like Canadian-born of European descent in BC, the result is less than 1 (in this case, about 0.64). If I restricted observations to the three major parties, then isn't it necessary that within each respective group the total vote share adds up to 1? Why is it that for some groups this isn't the case? Or am I just interpreting this incorrectly?

[MD Mubtasim-Fuad]

Nevermind I can't count apparently, please let me know though if anything looks suspicious