aboutsummaryrefslogtreecommitdiffstats
path: root/sql/dump_reingest_old.sql
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@archive.org>2022-10-03 10:16:26 -0700
committerBryan Newbold <bnewbold@archive.org>2022-10-03 10:16:26 -0700
commit54e14814080d9a706ff6f15694b3b54918200169 (patch)
treec2ce152acb64a365408a57728c18d960b954c1f5 /sql/dump_reingest_old.sql
parenta04468041cd81ad90aa76ec15788a5ffacb6eec2 (diff)
downloadsandcrawler-54e14814080d9a706ff6f15694b3b54918200169.tar.gz
sandcrawler-54e14814080d9a706ff6f15694b3b54918200169.zip
reingests: update scripts and SQL
Diffstat (limited to 'sql/dump_reingest_old.sql')
-rw-r--r--sql/dump_reingest_old.sql36
1 files changed, 36 insertions, 0 deletions
diff --git a/sql/dump_reingest_old.sql b/sql/dump_reingest_old.sql
new file mode 100644
index 0000000..7473420
--- /dev/null
+++ b/sql/dump_reingest_old.sql
@@ -0,0 +1,36 @@
+
+BEGIN TRANSACTION ISOLATION LEVEL SERIALIZABLE READ ONLY DEFERRABLE;
+
+COPY (
+ SELECT row_to_json(ingest_request.*) FROM ingest_request
+ LEFT JOIN ingest_file_result ON
+ ingest_file_result.base_url = ingest_request.base_url
+ AND ingest_file_result.ingest_type = ingest_request.ingest_type
+ WHERE
+ ingest_file_result.hit = false
+ AND ingest_request.created < NOW() - '6 day'::INTERVAL
+ -- AND ingest_request.created > NOW() - '181 day'::INTERVAL
+ AND (ingest_request.ingest_request_source = 'fatcat-changelog'
+ OR ingest_request.ingest_request_source = 'fatcat-ingest'
+ OR ingest_request.ingest_request_source = 'fatcat-ingest-container'
+ OR ingest_request.ingest_request_source = 'unpaywall'
+ OR ingest_request.ingest_request_source = 'arxiv'
+ OR ingest_request.ingest_request_source = 'pmc'
+ OR ingest_request.ingest_request_source = 'doaj'
+ OR ingest_request.ingest_request_source = 'dblp')
+ AND (
+ ingest_file_result.status like 'spn2-%'
+ -- OR ingest_file_result.status like 'no-capture'
+ -- OR ingest_file_result.status like 'cdx-error'
+ -- OR ingest_file_result.status like 'petabox-error'
+ )
+ AND ingest_file_result.status != 'spn2-error:invalid-url-syntax'
+ AND ingest_file_result.status != 'spn2-error:filesize-limit'
+ AND ingest_file_result.status != 'spn2-error:not-found'
+ AND ingest_file_result.status != 'spn2-error:blocked-url'
+ AND ingest_file_result.status != 'spn2-error:too-many-redirects'
+ AND ingest_file_result.status != 'spn2-error:network-authentication-required'
+ AND ingest_file_result.status != 'spn2-error:unknown'
+) TO '/srv/sandcrawler/tasks/reingest_old_current.rows.json';
+
+ROLLBACK;