aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
-rw-r--r--python/fatcat_tools/importers/ingest.py4
1 files changed, 3 insertions, 1 deletions
diff --git a/python/fatcat_tools/importers/ingest.py b/python/fatcat_tools/importers/ingest.py
index 7a98775d..04ff8db6 100644
--- a/python/fatcat_tools/importers/ingest.py
+++ b/python/fatcat_tools/importers/ingest.py
@@ -437,7 +437,9 @@ class IngestWebResultImporter(IngestFileResultImporter):
if 'revisit_cdx' in row:
terminal_cdx = row['revisit_cdx']
assert terminal_cdx['surt']
- assert terminal_cdx['url'] == terminal['terminal_url']
+ if terminal_cdx['url'] != terminal['terminal_url']:
+ self.counts['skip-terminal-url-mismatch'] += 1
+ return None
wc_cdx = []
# primary resource first