Make maintain_901() more robust in the presence of namespaces

Some records manage to get XML namespace declarations into each element,
like <datafield xmlns:marc="http://www.loc.gov/MARC21/slim"

This broke the old maintain_901(), so we'll make the regex more robust

git-svn-id: svn://svn.open-ils.org/ILS/trunk@16919 dcc99617-32d9-48b4-a31d-7c20da2025e4

diff --git a/Open-ILS/src/sql/Pg/002.functions.config.sql b/Open-ILS/src/sql/Pg/002.functions.config.sql
index 8f20de9..fb1d3b2 100644 (file)
--- a/Open-ILS/src/sql/Pg/002.functions.config.sql
+++ b/Open-ILS/src/sql/Pg/002.functions.config.sql
@@ -417,7 +417,8 @@ $f$ LANGUAGE PLPERLU;
 
 CREATE OR REPLACE FUNCTION maintain_901 () RETURNS TRIGGER AS $func$
 BEGIN
-    NEW.marc := REGEXP_REPLACE(NEW.marc, E'<datafield tag="901".+?</datafield>', '', 'g');
+    -- Remove any existing 901 fields before we insert the authoritative one
+    NEW.marc := REGEXP_REPLACE(NEW.marc, E'<datafield\s*[^<>]*?\s*tag="901".+?</datafield>', '', 'g');
     IF TG_TABLE_SCHEMA = 'biblio' THEN
         NEW.marc := REGEXP_REPLACE(
             NEW.marc,

diff --git a/Open-ILS/src/sql/Pg/002.schema.config.sql b/Open-ILS/src/sql/Pg/002.schema.config.sql
index ccdfc80..ecf1979 100644 (file)
--- a/Open-ILS/src/sql/Pg/002.schema.config.sql
+++ b/Open-ILS/src/sql/Pg/002.schema.config.sql
@@ -68,7 +68,7 @@ CREATE TABLE config.upgrade_log (
     install_date    TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT NOW()
 );
 
-INSERT INTO config.upgrade_log (version) VALUES ('0334'); -- Scott McKellar
+INSERT INTO config.upgrade_log (version) VALUES ('0335'); -- dbs
 
 CREATE TABLE config.bib_source (
        id              SERIAL  PRIMARY KEY,

diff --git a/Open-ILS/src/sql/Pg/upgrade/0335.schema.maintain_901_xmlns.sql b/Open-ILS/src/sql/Pg/upgrade/0335.schema.maintain_901_xmlns.sql
new file mode 100644 (file)
index 0000000..0fa4fd6
--- /dev/null
+++ b/Open-ILS/src/sql/Pg/upgrade/0335.schema.maintain_901_xmlns.sql
@@ -0,0 +1,63 @@
+BEGIN;
+
+-- Some records manage to get XML namespace declarations into each element,
+-- like <datafield xmlns:marc="http://www.loc.gov/MARC21/slim"
+-- This broke the old maintain_901(), so we'll make the regex more robust
+
+INSERT INTO config.upgrade_log (version) VALUES ('0335'); -- dbs
+
+CREATE OR REPLACE FUNCTION maintain_901 () RETURNS TRIGGER AS $func$
+BEGIN
+    -- Remove any existing 901 fields before we insert the authoritative one
+    NEW.marc := REGEXP_REPLACE(NEW.marc, E'<datafield\s*[^<>]*?\s*tag="901".+?</datafield>', '', 'g');
+    IF TG_TABLE_SCHEMA = 'biblio' THEN
+        NEW.marc := REGEXP_REPLACE(
+            NEW.marc,
+            E'(</(?:[^:]*?:)?record>)',
+            E'<datafield tag="901" ind1=" " ind2=" ">' ||
+                '<subfield code="a">' || NEW.tcn_value || E'</subfield>' ||
+                '<subfield code="b">' || NEW.tcn_source || E'</subfield>' ||
+                '<subfield code="c">' || NEW.id || E'</subfield>' ||
+                '<subfield code="t">' || TG_TABLE_SCHEMA || E'</subfield>' ||
+                CASE WHEN NEW.owner IS NOT NULL THEN '<subfield code="o">' || NEW.owner || E'</subfield>' ELSE '' END ||
+                CASE WHEN NEW.share_depth IS NOT NULL THEN '<subfield code="d">' || NEW.share_depth || E'</subfield>' ELSE '' END ||
+             E'</datafield>\\1'
+        );
+    ELSIF TG_TABLE_SCHEMA = 'authority' THEN
+        NEW.marc := REGEXP_REPLACE(
+            NEW.marc,
+            E'(</(?:[^:]*?:)?record>)',
+            E'<datafield tag="901" ind1=" " ind2=" ">' ||
+                '<subfield code="a">' || NEW.arn_value || E'</subfield>' ||
+                '<subfield code="b">' || NEW.arn_source || E'</subfield>' ||
+                '<subfield code="c">' || NEW.id || E'</subfield>' ||
+                '<subfield code="t">' || TG_TABLE_SCHEMA || E'</subfield>' ||
+             E'</datafield>\\1'
+        );
+    ELSIF TG_TABLE_SCHEMA = 'serial' THEN
+        NEW.marc := REGEXP_REPLACE(
+            NEW.marc,
+            E'(</(?:[^:]*?:)?record>)',
+            E'<datafield tag="901" ind1=" " ind2=" ">' ||
+                '<subfield code="c">' || NEW.id || E'</subfield>' ||
+                '<subfield code="t">' || TG_TABLE_SCHEMA || E'</subfield>' ||
+                '<subfield code="o">' || NEW.owning_lib || E'</subfield>' ||
+                CASE WHEN NEW.record IS NOT NULL THEN '<subfield code="r">' || NEW.record || E'</subfield>' ELSE '' END ||
+             E'</datafield>\\1'
+        );
+    ELSE
+        NEW.marc := REGEXP_REPLACE(
+            NEW.marc,
+            E'(</(?:[^:]*?:)?record>)',
+            E'<datafield tag="901" ind1=" " ind2=" ">' ||
+                '<subfield code="c">' || NEW.id || E'</subfield>' ||
+                '<subfield code="t">' || TG_TABLE_SCHEMA || E'</subfield>' ||
+             E'</datafield>\\1'
+        );
+    END IF;
+
+    RETURN NEW;
+END;
+$func$ LANGUAGE PLPGSQL;
+
+COMMIT;