aboutsummaryrefslogtreecommitdiffstats
path: root/sql/dump_unextracted_pdf.sql
diff options
context:
space:
mode:
Diffstat (limited to 'sql/dump_unextracted_pdf.sql')
-rw-r--r--sql/dump_unextracted_pdf.sql14
1 files changed, 10 insertions, 4 deletions
diff --git a/sql/dump_unextracted_pdf.sql b/sql/dump_unextracted_pdf.sql
index 4d264b9..a7fb920 100644
--- a/sql/dump_unextracted_pdf.sql
+++ b/sql/dump_unextracted_pdf.sql
@@ -1,16 +1,22 @@
-- Run like:
--- psql sandcrawler < dump_unextracted_pdf.sql > dump_unextracted_pdf.2019-09-23.json
+-- psql sandcrawler < dump_unextracted_pdf.sql
BEGIN TRANSACTION ISOLATION LEVEL SERIALIZABLE READ ONLY DEFERRABLE;
COPY (
- SELECT row_to_json(cdx)
+ SELECT DISTINCT ON (cdx.sha1hex) row_to_json(cdx)
FROM grobid
LEFT JOIN cdx ON grobid.sha1hex = cdx.sha1hex
+ --LEFT JOIN fatcat_file ON grobid.sha1hex = fatcat_file.sha1hex
+ LEFT JOIN ingest_file_result ON grobid.sha1hex = ingest_file_result.terminal_sha1hex
+ LEFT JOIN pdf_meta ON grobid.sha1hex = pdf_meta.sha1hex
+ WHERE cdx.sha1hex IS NOT NULL
+ --AND fatcat_file.sha1hex IS NOT NULL
+ AND ingest_file_result.terminal_sha1hex IS NOT NULL
+ AND pdf_meta.sha1hex IS NULL
)
--- TO '/grande/snapshots/dump_unextracted_pdf.2020-06-25.json';
-TO STDOUT
+TO '/srv/sandcrawler/tasks/dump_unextracted_pdf.ingest.2020-10-21.json'
WITH NULL '';
ROLLBACK;