Fetch existing keys in parallelarchive/prototype/fdb-layer-view-tracingprototype/fdb-layer-view-tracing

Fetching the existing keys was a time sink when updating the indices.
This moves the key fetching into a single pass for all docs
simultaneously.

2 files changed, 31 insertions, 19 deletions

diff --git a/src/couch_views/src/couch_views_fdb.erl b/src/couch_views/src/couch_views_fdb.erl
index b85f2bd18..482d35fcf 100644
--- a/src/couch_views/src/couch_views_fdb.erl
+++ b/src/couch_views/src/couch_views_fdb.erl
@@ -21,6 +21,7 @@
 
     fold_map_idx/6,
 
+    get_existing_keys/3,
     write_doc/4
 ]).
 
@@ -107,13 +108,34 @@ fold_map_idx(TxDb, Sig, ViewId, Options, Callback, Acc0) ->
     Acc1.
 
 
+get_existing_keys(TxDb, Sig, Docs) ->
+    #{
+        tx := Tx,
+        db_prefix := DbPrefix
+    } = TxDb,
+
+    Futures = lists:map(fun(#{id := DocId}) ->
+        {Start, End} = id_idx_range(DbPrefix, Sig, DocId),
+        erlfdb:fold_range_future(Tx, Start, End, [])
+    end, Docs),
+
+    lists:zipwith(fun(Future, #{id := DocId} = Doc) ->
+        Entries = erlfdb:fold_range_wait(Tx, Future, fun({K, V}, Acc) ->
+            {?DB_VIEWS, ?VIEW_DATA, Sig, ?VIEW_ID_RANGE, DocId, ViewId} =
+                    erlfdb_tuple:unpack(K, DbPrefix),
+            [TotalKeys, TotalSize, UniqueKeys] = couch_views_encoding:decode(V),
+            [{ViewId, TotalKeys, TotalSize, UniqueKeys} | Acc]
+        end, []),
+        Doc#{existing_keys => lists:sort(Entries)}
+    end, Futures, Docs).
+
+
 write_doc(TxDb, Sig, _ViewIds, #{deleted := true} = Doc) ->
     #{
-        id := DocId
+        id := DocId,
+        existing_keys := ExistingViewKeys
     } = Doc,
 
-    ExistingViewKeys = get_view_keys(TxDb, Sig, DocId),
-
     clear_id_idx(TxDb, Sig, DocId),
     lists:foreach(fun({ViewId, TotalKeys, TotalSize, UniqueKeys}) ->
         clear_map_idx(TxDb, Sig, ViewId, DocId, UniqueKeys),
@@ -124,11 +146,11 @@ write_doc(TxDb, Sig, _ViewIds, #{deleted := true} = Doc) ->
 write_doc(TxDb, Sig, ViewIds, Doc) ->
     #{
         id := DocId,
+        existing_keys := ExistingViewKeys,
         results := Results,
         kv_sizes := KVSizes
     } = Doc,
 
-    ExistingViewKeys = get_view_keys(TxDb, Sig, DocId),
 
     clear_id_idx(TxDb, Sig, DocId),
     lists:foreach(fun({ViewId, NewRows, KVSize}) ->
@@ -246,20 +268,6 @@ update_map_idx(TxDb, Sig, ViewId, DocId, ExistingKeys, NewRows) ->
     end, KVsToAdd).
 
 
-get_view_keys(TxDb, Sig, DocId) ->
-    #{
-        tx := Tx,
-        db_prefix := DbPrefix
-    } = TxDb,
-    {Start, End} = id_idx_range(DbPrefix, Sig, DocId),
-    lists:map(fun({K, V}) ->
-        {?DB_VIEWS, ?VIEW_DATA, Sig, ?VIEW_ID_RANGE, DocId, ViewId} =
-                erlfdb_tuple:unpack(K, DbPrefix),
-        [TotalKeys, TotalSize, UniqueKeys] = couch_views_encoding:decode(V),
-        {ViewId, TotalKeys, TotalSize, UniqueKeys}
-    end, erlfdb:get_range(Tx, Start, End, [])).
-
-
 update_row_count(TxDb, Sig, ViewId, Increment) ->
     #{
         tx := Tx,
diff --git a/src/couch_views/src/couch_views_indexer.erl b/src/couch_views/src/couch_views_indexer.erl
index b787574d0..47860051f 100644
--- a/src/couch_views/src/couch_views_indexer.erl
+++ b/src/couch_views/src/couch_views_indexer.erl
@@ -310,7 +310,7 @@ map_docs(Mrst, Docs) ->
     {Mrst1, MappedDocs}.
 
 
-write_docs(TxDb, Mrst, Docs, State) ->
+write_docs(TxDb, Mrst, Docs0, State) ->
     #mrst{
         views = Views,
         sig = Sig
@@ -324,6 +324,10 @@ write_docs(TxDb, Mrst, Docs, State) ->
     KeyLimit = key_size_limit(),
     ValLimit = value_size_limit(),
 
+    Docs = ctrace:with_span(get_existing_keys, fun() ->
+        couch_views_fdb:get_existing_keys(TxDb, Sig, Docs0)
+    end),
+
     lists:foreach(fun(Doc0) ->
         Doc1 = calculate_kv_sizes(Mrst, Doc0, KeyLimit, ValLimit),
         Tags = #{doc_id => maps:get(id, Doc0)},