Deduplicate builds and add a unique index

Duplicate builds could creep in if the code to create them ran concurrently. I
didn't exclude them initially, as I was unsure if there should be such a
restriction, but at least for now, Cuirass builds map exactly to a single
derivation, so use the same restriction here.

4 files changed, 48 insertions, 0 deletions

diff --git a/sqitch/deploy/sort_out_duplicate_builds.sql b/sqitch/deploy/sort_out_duplicate_builds.sql
new file mode 100644
index 0000000..92bd2c5
--- /dev/null
+++ b/sqitch/deploy/sort_out_duplicate_builds.sql
@@ -0,0 +1,33 @@
+-- Deploy guix-data-service:sort_out_duplicate_builds to pg
+
+BEGIN;
+
+DELETE FROM build_status WHERE build_id IN (
+  SELECT builds.id
+  FROM builds
+  INNER JOIN (
+    SELECT derivation_file_name, MIN(id) AS id
+    FROM builds
+    GROUP BY build_server_id, derivation_file_name
+    HAVING COUNT(DISTINCT id) > 1
+  ) AS min_ids
+    ON min_ids.derivation_file_name = builds.derivation_file_name AND
+       min_ids.id != builds.id
+);
+
+DELETE FROM builds WHERE id IN (
+  SELECT builds.id
+  FROM builds
+  INNER JOIN (
+    SELECT derivation_file_name, MIN(id) AS id
+    FROM builds
+    GROUP BY build_server_id, derivation_file_name
+    HAVING COUNT(DISTINCT id) > 1
+  ) AS min_ids
+    ON min_ids.derivation_file_name = builds.derivation_file_name AND
+       min_ids.id != builds.id
+);
+
+CREATE UNIQUE INDEX ON builds (build_server_id, derivation_file_name);
+
+COMMIT;
diff --git a/sqitch/revert/sort_out_duplicate_builds.sql b/sqitch/revert/sort_out_duplicate_builds.sql
new file mode 100644
index 0000000..153a4ef
--- /dev/null
+++ b/sqitch/revert/sort_out_duplicate_builds.sql
@@ -0,0 +1,7 @@
+-- Revert guix-data-service:sort_out_duplicate_builds from pg
+
+BEGIN;
+
+DROP INDEX builds_build_server_id_derivation_file_name_idx;
+
+COMMIT;
diff --git a/sqitch/sqitch.plan b/sqitch/sqitch.plan
index ee65c61..c323eef 100644
--- a/sqitch/sqitch.plan
+++ b/sqitch/sqitch.plan
@@ -34,3 +34,4 @@ add_index_on_builds_derivation_file_name 2019-12-03T16:23:55Z   <chris@phact> #
 narinfo_fetch_record 2019-12-03T20:53:28Z   <chris@phact> # Add a table to describe fetching a narinfo
 change_nars_size_to_bigint 2019-12-04T21:24:21Z   <chris@phact> # Change nars.size to bigint
 change_nar_urls_size_to_bigint 2019-12-04T21:49:07Z   <chris@phact> # Change nar_urls.size to bigint
+sort_out_duplicate_builds 2019-12-05T12:43:53Z   <chris@phact> # Sort out duplicate builds
diff --git a/sqitch/verify/sort_out_duplicate_builds.sql b/sqitch/verify/sort_out_duplicate_builds.sql
new file mode 100644
index 0000000..a63cda4
--- /dev/null
+++ b/sqitch/verify/sort_out_duplicate_builds.sql
@@ -0,0 +1,7 @@
+-- Verify guix-data-service:sort_out_duplicate_builds on pg
+
+BEGIN;
+
+-- XXX Add verifications here.
+
+ROLLBACK;