DedupeTool: Use matchType=exact for CDX server

iipc · Sep 23, 2024 · 5a9b5ff · 5a9b5ff
1 parent a22e1f9
commit 5a9b5ff
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/src/org/netpreserve/jwarc/tools/DedupeTool.java b/src/org/netpreserve/jwarc/tools/DedupeTool.java
@@ -65,7 +65,7 @@ private WarcRevisit deduplicate(WarcRecord record) throws IOException {
     }
 
     private CdxRecord findMatchingRecord(WarcCaptureRecord capture, String digest) throws IOException {
-        URL queryUrl = new URL(cdxServer + "?sort=reverse&rows=10&url=" + URLEncoder.encode(capture.target(), UTF_8.name()));
+        URL queryUrl = new URL(cdxServer + "?sort=reverse&rows=10&matchType=exact&url=" + URLEncoder.encode(capture.target(), UTF_8.name()));
         try (CdxReader response = new CdxReader(queryUrl.openStream())) {
             for (CdxRecord record : response) {
                 if (digest.equalsIgnoreCase(record.digest())) {