docs: clarify mapping data update workflow

Author: zonble

.github/instructions/Data.instructions.md

@@ -6,11 +6,19 @@ These files define the mapping of phrases and words to their Bopomofo representa
 
 In most cases, developers will add new Chinese characters or phrases rather than delete existing ones. When adding new characters or phrases, please perform the following checks:
 
-## Heterophonic Characters
+## Adding Phrases 
 
-When adding a new entry—such as a character with a Bopomofo reading—into `BPMFMappings.txt`, check if there is already an entry for the same character with a different Bopomofo reading. If so, this indicates a heterophonic character, and you should review the frequency of each reading.
+When adding a new phrase containing multiple characters to `BPMFMappings.txt`, ensure that `phrase.occ` is also updated with the same phrase and its frequency. The frequency should be a positive integer.
 
-Often, the new reading will be much less common than the existing one. In such cases, add a comment to indicate that this is a heterophonic character and note the frequency of each reading.
+## Sorting
+
+Both `BPMFMappings.txt` and `phrase.occ` must stay sorted using the C locale. After making changes, run `LC_ALL=c sort -o BPMFMappings.txt BPMFMappings.txt` and `LC_ALL=c sort -o phrase.occ phrase.occ` before committing.
+
+## Heterophony Characters
+
+When adding a new entry—such as a character with a Bopomofo reading—into `BPMFMappings.txt`, check if there is already an entry for the same character with a different Bopomofo reading. If so, this indicates a heterophony character, and you should review the frequency of each reading.
+
+Often, the new reading will be much less common than the existing one. In such cases, add a comment to indicate that this is a heterophony character and note the frequency of each reading.
 
 To reflect the lower frequency of the new reading, place the default reading in `heterophony1.list` and the new reading in `heterophony2.list`, and so on.