Why "sort" behaves weirdly when "LANG" is not set?

Question

In my bash, "LANG" was set to "en_us.UTF-8", and "sort" works fine. But if I unset "LANG", then "sort" works weirdly. Even with "LC_ALL=C" or "LC_ALL=POSIX". Anyone know why "sort" doesn't work when "LANG" is not set?

$ echo $LANG
en_US.UTF-8

$ sort -t$'\t' -k1,3 -gr
-4.445905   1   0.965933
-4.445905   1   0.76126
-4.445905   1   0.74816
-4.445905   1   0.633251
-4.445905   1   0.596921
-4.445905   1   0.464271
-4.445905   1   0.406553
-4.445905   1   0.350968
-4.445905   1   0.307701
-4.445905   1   0.188411
-4.445905   1   0.0377354
-4.445905   1   0.0221152
-4.445905   -1  0.999663
-4.445905   -1  0.987287
-4.445905   -1  0.97882
-4.445905   -1  0.969835
-4.445905   -1  0.96705
-4.445905   -1  0.964982
-4.445905   -1  0.920791
-4.445905   -1  0.901001
-4.445905   -1  0.877351
-4.445905   -1  0.87702

-4.445905   -1  0.999663
-4.445905   -1  0.987287
-4.445905   -1  0.97882
-4.445905   -1  0.969835
-4.445905   -1  0.96705
-4.445905   1   0.965933
-4.445905   -1  0.964982
-4.445905   -1  0.920791
-4.445905   -1  0.901001
-4.445905   -1  0.877351
-4.445905   -1  0.87702
-4.445905   1   0.76126
-4.445905   1   0.74816
-4.445905   1   0.633251
-4.445905   1   0.596921
-4.445905   1   0.464271
-4.445905   1   0.406553
-4.445905   1   0.350968
-4.445905   1   0.307701
-4.445905   1   0.188411
-4.445905   1   0.0377354
-4.445905   1   0.0221152

$ unset LANG

$ sort -t$'\t' -k1,3 -gr
-4.445905   1   0.965933
-4.445905   1   0.76126
-4.445905   1   0.74816
-4.445905   1   0.633251
-4.445905   1   0.596921
-4.445905   1   0.464271
-4.445905   1   0.406553
-4.445905   1   0.350968
-4.445905   1   0.307701
-4.445905   1   0.188411
-4.445905   1   0.0377354
-4.445905   1   0.0221152
-4.445905   -1  0.999663
-4.445905   -1  0.987287
-4.445905   -1  0.97882
-4.445905   -1  0.969835
-4.445905   -1  0.96705
-4.445905   -1  0.964982
-4.445905   -1  0.920791
-4.445905   -1  0.901001
-4.445905   -1  0.877351
-4.445905   -1  0.87702

-4.445905   1   0.965933
-4.445905   1   0.76126
-4.445905   1   0.74816
-4.445905   1   0.633251
-4.445905   1   0.596921
-4.445905   1   0.464271
-4.445905   1   0.406553
-4.445905   1   0.350968
-4.445905   1   0.307701
-4.445905   1   0.188411
-4.445905   1   0.0377354
-4.445905   1   0.0221152
-4.445905   -1  0.999663
-4.445905   -1  0.987287
-4.445905   -1  0.97882
-4.445905   -1  0.969835
-4.445905   -1  0.96705
-4.445905   -1  0.964982
-4.445905   -1  0.920791
-4.445905   -1  0.901001
-4.445905   -1  0.877351
-4.445905   -1  0.87702

$ LC_ALL=POSIX sort -t$'\t' -k1,3 -gr
-4.445905   1   0.965933
-4.445905   1   0.76126
-4.445905   1   0.74816
-4.445905   1   0.633251
-4.445905   1   0.596921
-4.445905   1   0.464271
-4.445905   1   0.406553
-4.445905   1   0.350968
-4.445905   1   0.307701
-4.445905   1   0.188411
-4.445905   1   0.0377354
-4.445905   1   0.0221152
-4.445905   -1  0.999663
-4.445905   -1  0.987287
-4.445905   -1  0.97882
-4.445905   -1  0.969835
-4.445905   -1  0.96705
-4.445905   -1  0.964982
-4.445905   -1  0.920791
-4.445905   -1  0.901001
-4.445905   -1  0.877351
-4.445905   -1  0.87702

-4.445905   1   0.965933
-4.445905   1   0.76126
-4.445905   1   0.74816
-4.445905   1   0.633251
-4.445905   1   0.596921
-4.445905   1   0.464271
-4.445905   1   0.406553
-4.445905   1   0.350968
-4.445905   1   0.307701
-4.445905   1   0.188411
-4.445905   1   0.0377354
-4.445905   1   0.0221152
-4.445905   -1  0.999663
-4.445905   -1  0.987287
-4.445905   -1  0.97882
-4.445905   -1  0.969835
-4.445905   -1  0.96705
-4.445905   -1  0.964982
-4.445905   -1  0.920791
-4.445905   -1  0.901001
-4.445905   -1  0.877351
-4.445905   -1  0.87702

"LANG" was set to "en_us.UTF-8", and "sort" works fine. But if I unset "LANG", then "sort" works weirdly. Isn't it the exact opposite way around? — n. m. could be an AI
– n. m. could be an AI, Commented Jan 31, 2015 at 8:45
The first quoted sort does sort the first then the third column descending, which is expected. The later two "sort" after "unset LANG" doesn't sort the third column but instead sort the second column, which is unexpected. Am I missing something? — Roy
– Roy, Commented Jan 31, 2015 at 8:50
Oh i see the key is actually "from POS1 to POS2 inclusive", so "-k 1,3" means sort using key column1+column2+column3. n.m. you were correct, the later two sort actually did the right thing. — Roy
– Roy, Commented Jan 31, 2015 at 8:52
-k1,3 means keys are from the first to the third column inclusively. If you want to ignore the second column, use -k1 -k3. — n. m. could be an AI
– n. m. could be an AI, Commented Jan 31, 2015 at 8:55

Roy · Accepted Answer · 2015-01-31 08:54:12Z

1

Apparently I didn't correctly understand the sort key option "-k". It is actually "from POS1 to POS2 inclusive", so "-k 1,3" means sort using key column1+column2+column3. Thus the later two "sort" are doing the correct thing, while the first is actually weird.

answered Jan 31, 2015 at 8:54

Roy

9452 gold badges13 silver badges29 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Why "sort" behaves weirdly when "LANG" is not set?

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related