Remove unnecessary sorts before joins in ztests

Remove sort operators before joins in ztests that are unnecessary now
that join automatically sorts its inputs.

compiler/ztests/join-desc.yaml is left unmodified since it is explicitly
testing inputs with descending order.

zfmt/ztests/decls.yaml is modified for consistency, even though removing
the sorts affects the output (which is formatted Zed, not query
results).

compiler/ztests/from-pass.yaml

@@ -4,8 +4,8 @@ inputs:
   - name: join.zed
     data: |
       * | from (
-        pass => sort x
-        file right.zson => sort y
+        pass
+        file right.zson
       ) | inner join on x=y matched:=true
   - name: left.zson
     data: |
@@ -15,6 +15,7 @@ inputs:
   - name: right.zson
     data: |
       {y:2,y:"y-two"}
+
 outputs:
   - name: stdout
     data: |

docs/language/ztests/language-directed-acyclic-flow-graphs-2.yaml

@@ -43,8 +43,8 @@ inputs:
   - name: join.zed
     data: |
       from (
-        pool PoolOne => put color:=upper(color) | put namelen:=len(key) | sort key
-        pool PoolTwo => put price:=price+0.5 | put tag:="mytag" | sort key
+        pool PoolOne => put color:=upper(color) | put namelen:=len(key)
+        pool PoolTwo => put price:=price+0.5 | put tag:="mytag"
       ) | join on key=key priceinfo:=this | drop priceinfo.key
 
 outputs:

lake/ztests/cross-pool-join.yaml

@@ -24,8 +24,8 @@ inputs:
   - name: join.zed
     data: |
       from (
-        pool fruit => cut f:=this | sort f.flavor
-        pool person => cut p:=this | sort p.likes
+        pool fruit => cut f:=this
+        pool person => cut p:=this
       )
       | join on f.flavor=p.likes p:=p
       | summarize likes:=union(f.name) by name:=p.name | sort name

lake/ztests/index/apply.yaml

@@ -18,8 +18,8 @@ inputs:
   - name: query.zed
     data: |
       from (
-        pool :index_rules => sort id
-        pool test@main:indexes => sort rule.id | cut o:=this
+        pool :index_rules
+        pool test@main:indexes => cut o:=this
       )
       | left join on id = o.rule.id o
       | count(o) by name, fields

lake/ztests/index/update.yaml

@@ -23,8 +23,8 @@ inputs:
   - name: query.zed
     data: |
       from (
-        pool :index_rules => sort id
-        pool test@main:indexes => sort rule.id | cut o:=this
+        pool :index_rules
+        pool test@main:indexes => cut o:=this
       )
       | left join on id = o.rule.id o
       | count(o) by name,fields

runtime/op/join/ztests/cut-like.yaml

@@ -4,8 +4,8 @@ inputs:
   - name: test.zed
     data: |
       from (
-        file people.ndjson => sort likes
-        file fruit.ndjson => sort flavor
+        file people.ndjson
+        file fruit.ndjson
       ) | left join on likes=flavor fruit:=name,note:=quiet(note)
   - name: people.ndjson
     data: |

runtime/op/join/ztests/kinds.yaml

@@ -1,12 +1,12 @@
 script: |
   echo === ANTI ===
-  zq -z 'anti join (file B.zson) on a=b | sort a' A.zson
+  zq -z 'anti join (file B.zson) on a=b' A.zson
   echo === LEFT ===
-  zq -z 'left join (file B.zson) on a=b hit:=sb | sort a' A.zson
+  zq -z 'left join (file B.zson) on a=b hit:=sb' A.zson
   echo === INNER ===
-  zq -z 'inner join (file B.zson) on a=b hit:=sb | sort a' A.zson
+  zq -z 'inner join (file B.zson) on a=b hit:=sb' A.zson
   echo === RIGHT ===
-  zq -z 'right join (file C.zson) on b=c hit:=sb | sort c' B.zson
+  zq -z 'right join (file C.zson) on b=c hit:=sb' B.zson
 
 inputs:
   - name: A.zson

service/ztests/index/apply.yaml

@@ -19,8 +19,8 @@ inputs:
   - name: query.zed
     data: |
       from (
-        pool :index_rules => sort id
-        pool test@main:indexes => sort rule.id | cut o:=this
+        pool :index_rules
+        pool test@main:indexes => cut o:=this
       )
       | left join on id = o.rule.id o
       | count(o) by name, fields

service/ztests/index/update.yaml

@@ -24,8 +24,8 @@ inputs:
   - name: query.zed
     data: |
       from (
-        pool :index_rules => sort id
-        pool test@main:indexes => sort rule.id | cut o:=this
+        pool :index_rules
+        pool test@main:indexes => cut o:=this
       )
       | left join on id = o.rule.id o
       | count(o) by name,fields

zfmt/ztests/decls.yaml

@@ -13,9 +13,9 @@ inputs:
       op stamp(assignee): ( yield {...this, assignee, ts: now()} )
       op nop(foo): ( pass )
       op joinTest(left_file, right_file, left_key, right_key, left_dest, right_source): (
-        file left_file | sort left_key
+        file left_file
         | inner join (
-          file right_file | sort right_key
+          file right_file
         ) on left_key = right_key left_dest := right_source
       )
       joinTest("fruit.ndjson", "people.ndjson", flavor, likes, eater, name)
@@ -42,12 +42,10 @@ outputs:
         from (
           file left_file
         )
-        | sort left_key
         | join (
           from (
             file right_file
           )
-          | sort right_key
         ) on left_key=right_key left_dest:=right_source
       )
       joinTest("fruit.ndjson", "people.ndjson", flavor, likes, eater, name)
@@ -64,13 +62,11 @@ outputs:
         )
         
         file fruit.ndjson
-        | sort flavor
         | fork (
           =>
             pass
           =>
             file people.ndjson
-            | sort likes
         )
         | join on flavor=likes eater:=name
         | yield {...this,assignee:"bob",ts:now()}