ALS-6511: Fix avro formatting issue

hms-dbmi · Jul 22, 2024 · bc1cda5 · bc1cda5
1 parent 1a3fd09
commit bc1cda5
Show file tree

Hide file tree

Showing 2 changed files with 27 additions and 2 deletions.
diff --git a/processing/src/main/java/edu/harvard/hms/dbmi/avillach/hpds/processing/io/PfbWriter.java b/processing/src/main/java/edu/harvard/hms/dbmi/avillach/hpds/processing/io/PfbWriter.java
@@ -73,8 +73,12 @@ public void writeHeader(String[] data) {
         writeMetadata();
     }
 
-    private String formatFieldName(String s) {
-        return s.replaceAll("\\\\", "_");
+    protected String formatFieldName(String s) {
+        String formattedFieldName = s.replaceAll("\\W", "_");
+        if (Character.isDigit(formattedFieldName.charAt(0))) {
+            return "_" + formattedFieldName;
+        }
+        return formattedFieldName;
     }
 
     private void writeMetadata() {

diff --git a/processing/src/test/java/edu/harvard/hms/dbmi/avillach/hpds/processing/io/PfbWriterTest.java b/processing/src/test/java/edu/harvard/hms/dbmi/avillach/hpds/processing/io/PfbWriterTest.java
@@ -24,4 +24,25 @@ public void writeValidPFB() {
         pfbWriter.close();
         // todo: validate this programatically
     }
+
+    @Test
+    public void formatFieldName_spacesAndBackslashes_replacedWithUnderscore() {
+        PfbWriter pfbWriter = new PfbWriter(new File("target/test-result.avro"));
+        String formattedName = pfbWriter.formatFieldName("\\Topmed Study Accession with Subject ID\\\\");
+        assertEquals("_Topmed_Study_Accession_with_Subject_ID__", formattedName);
+    }
+
+    @Test
+    public void formatFieldName_startsWithDigit_prependUnderscore() {
+        PfbWriter pfbWriter = new PfbWriter(new File("target/test-result.avro"));
+        String formattedName = pfbWriter.formatFieldName("123Topmed Study Accession with Subject ID\\\\");
+        assertEquals("_123Topmed_Study_Accession_with_Subject_ID__", formattedName);
+    }
+
+    @Test
+    public void formatFieldName_randomGarbage_replaceWithUnderscore() {
+        PfbWriter pfbWriter = new PfbWriter(new File("target/test-result.avro"));
+        String formattedName = pfbWriter.formatFieldName("$$$my garbage @vro var!able nam#");
+        assertEquals("___my_garbage__vro_var_able_nam_", formattedName);
+    }
 }