Fix #15

Emory-HITI · Apr 3, 2021 · d4478fe · d4478fe
2 parents 92e0d59 + 4e1a602
commit d4478fe
Showing 1 changed file with 19 additions and 14 deletions.
diff --git a/modules/cold-extraction/ColdDataRetriever.py b/modules/cold-extraction/ColdDataRetriever.py
@@ -131,23 +131,28 @@ def initialize():
     reader = csv.reader(f)
     next(f)
     for row in reader:
+        row = [x.strip() for x in row]
         if (extraction_type == 'empi_date'):
-            patients.append(row[patient_index])
-            temp_date = row[date_index]
-            dt_stamp = datetime.datetime.strptime(temp_date, date_format)
-            date_str = dt_stamp.strftime('%Y%m%d')
-            dates.append(date_str)
-            length = len(patients)
+            if not ((row[patient_index] == "") or (row[date_index] == "")):
+                patients.append(row[patient_index])
+                temp_date = row[date_index]
+                dt_stamp = datetime.datetime.strptime(temp_date, date_format)
+                date_str = dt_stamp.strftime('%Y%m%d')
+                dates.append(date_str)
+                length = len(patients)
         elif (extraction_type == 'empi'):
-            patients.append(row[patient_index])
-            length = len(patients)
+            if not ((row[patient_index] == "")):
+                patients.append(row[patient_index])
+                length = len(patients)
         elif (extraction_type == 'accession'):
-            accessions.append(row[accession_index])
-            length = len(accessions)
+            if not ((row[accession_index] == "")):
+                accessions.append(row[accession_index])
+                length = len(accessions)
         elif (extraction_type == 'empi_accession'):
-            patients.append(row[patient_index])
-            accessions.append(row[accession_index])
-            length = len(accessions)
+            if not ((row[patient_index] == "") or (row[accession_index] == "")):
+                patients.append(row[patient_index])
+                accessions.append(row[accession_index])
+                length = len(accessions)
 
 
 # Run the retrieval only once, when the extraction script starts, and keep it running in a separate thread.
@@ -277,4 +282,4 @@ def run_threaded(job_func):
         time.sleep(1)
     except KeyboardInterrupt:
         check_kill_process()
-        sys.exit(0)
+        sys.exit(0)