dump_regrobid_pdf_petabox.sql script

1 files changed, 15 insertions, 0 deletions

diff --git a/sql/dump_regrobid_pdf_petabox.sql b/sql/dump_regrobid_pdf_petabox.sql
new file mode 100644
index 0000000..3ca8085
--- /dev/null
+++ b/sql/dump_regrobid_pdf_petabox.sql
@@ -0,0 +1,15 @@
+
+-- Run like:
+--   psql sandcrawler < dump_regrobid_pdf_petabox.sql
+--   cat dump_regrobid_pdf_petabox.2020-02-03.json | sort -S 4G | uniq -w 40 | cut -f2 > dump_regrobid_pdf_petabox.2020-02-03.uniq.json
+
+BEGIN TRANSACTION ISOLATION LEVEL SERIALIZABLE READ ONLY DEFERRABLE;
+
+COPY (
+    SELECT petabox.sha1hex, row_to_json(petabox) FROM petabox
+    WHERE EXISTS (SELECT grobid.sha1hex FROM grobid WHERE petabox.sha1hex = grobid.sha1hex AND grobid.grobid_version IS NULL)
+)
+TO '/grande/snapshots/dump_regrobid_pdf_petabox.2020-02-03.json'
+WITH NULL '';
+
+ROLLBACK;