I am merging 2 data sets, in master data set, the variables are byte with no data, and in the using data set they are strings.
When they are merged, the data for these 2 variables is is lost.
I've tried this with and without the tostring and format with the same results.
This data set is expanding so I want to avoid having to check and manually input the missing data.
Thanks for any ideas.

Code:
Here is the data I want to preserve:
use "PlacesLived_V3.dta"

tab1 a_2_districtText1 a_2_VDC1

-> tabulation of a_2_districtText1  

a_2_distric |
     tText1 |      Freq.     Percent        Cum.
------------+-----------------------------------
 Bhairahawa |          1       50.00       50.00
    Chitwan |          1       50.00      100.00
------------+-----------------------------------
      Total |          2      100.00

-> tabulation of a_2_VDC1  

   a_2_VDC1 |      Freq.     Percent        Cum.
------------+-----------------------------------
 Bhairahawa |          1       50.00       50.00
      Tadhi |          1       50.00      100.00
------------+-----------------------------------
      Total |          2      100.00


This is the master data set:
use "ExploreV5", clear
tostring a_2_districtText1 a_2_VDC1, replace force 
format a_2_districtText1 a_2_VDC1 %11s

a_2_districtText1 was byte now str1
a_2_VDC1 was byte now str1


merge 1:1 id using PlacesLived_V3, force
(note: variable a_2_districtText1 was str1, now str10 to accommodate using data's values)
(note: variable a_2_VDC1 was str1, now str10 to accommodate using data's values)

tab1 a_2_districtText1 a_2_VDC1

-> tabulation of a_2_districtText1  
no observations

-> tabulation of a_2_VDC1  
no observations