Commit ad39be5d authored by R.W.Majeed's avatar R.W.Majeed
Browse files

test case for duplicate patient

parent c3bed170

histream-import/src/main/java/de/sekmi/histream/etl/validation/DuplicatePatientException.java

0 → 100644
+14 −0
Original line number Diff line number Diff line 
package de.sekmi.histream.etl.validation;



public class DuplicatePatientException extends ValidationException {



	/**

	 * 

	 */

	private static final long serialVersionUID = 1L;



	public DuplicatePatientException(String patientId) {

		super("Duplicate patient '"+patientId+"'");

	}



}

histream-import/src/main/java/de/sekmi/histream/etl/validation/Validator.java

+1 −1
Original line number Diff line number Diff line
@@ -43,7 +43,7 @@ public class Validator extends AbstractObservationHandler{ 
			prevPatient = patid; // remember patient to suppress errors for the same patient



			if( patients.contains(patid) ){

				throw new ValidationException("Duplicate patient '"+patid+"'");

				throw new DuplicatePatientException(patid);

			}else{

				patients.add(patid);

			}

histream-import/src/test/java/de/sekmi/histream/etl/validation/TestValidator.java

+14 −1
Original line number Diff line number Diff line 
package de.sekmi.histream.etl.validation;



import org.junit.Assert;

import org.junit.Test;



import de.sekmi.histream.ObservationSupplier;
@@ -17,4 +18,16 @@ public class TestValidator { 
			Streams.transfer(os, v);

		}

	}

	@Test

	public void validateData2() throws Exception{

		try( ObservationSupplier os = ETLObservationSupplier.load(getClass().getResource("/data/test-2-datasource.xml")) ){

			Validator v = new Validator();

			v.setErrorHandler(e -> {throw new RuntimeException(e);});

			Streams.transfer(os, v);

		}catch( RuntimeException e ){

			Assert.assertTrue(e.getCause() instanceof DuplicatePatientException);

			return;

		}

		Assert.fail("Exception expected");

	}

}

histream-import/src/test/resources/data/test-1-datasource.xml

+8 −1
Original line number Diff line number Diff line
@@ -68,7 +68,7 @@ 
				<start column="zeitpunkt" format="d.M.u[ H[:m[:s]]]"/>

				<unit constant-value="mmol/l" />

				<modifier id="other">

					<value xsi:type="string" constant-value=""/>

					<value xsi:type="string" na="" constant-value=""/>

				</modifier>

			</concept>

			<concept id="kalium">
@@ -104,6 +104,13 @@ 
			<start column="start_ts" format="u-M-d"/>

			<end column="end" format="u-M-d" na="@"/>

			<!-- location -->



			<!-- if group is non-na, all following eav rows with the same group value

			are considered facts in a single group. A group must contain a primary

			concept and any number of modifiers. All rows within the group must have

			the same patient,encounter and start timestamp. -->

			<!-- <group column="group" na="@" /> -->



			<type column="type"/>

			<value column="value" na="@"/>

			<unit column="unit" na="@"/>

histream-import/src/test/resources/data/test-2-datasource.xml

0 → 100644
+138 −0
Original line number Diff line number Diff line 
<?xml version="1.0" encoding="UTF-8"?>

<datasource version="1.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" >

	<meta>

		<id>test-1</id>

		<etl-strategy>replace-source</etl-strategy>

	</meta>

	

	<patient-table>

		<source xsi:type="plain-file">

			<url>test-2-patients-duplicate.txt</url>

			<separator>\t</separator>

		</source>

		<idat>

			<patient-id column="patid"/>

			<given-name column="vorname"/>

			<surname column="nachname"/>

			<birthdate format="d.M.u" na="" column="geburtsdatum"/>

			<deathdate format="d.M.u" na="" column="verstorben"/>

			<gender column="geschlecht" na="">

				<map> <!-- maps a column -->

					<case value="W" set-value="female"/>

					<case value="M" set-value="male"/>

					<case value="X" set-value="indeterminate"/>

					<otherwise set-value="" log-warning="Unexpected gender value"/>

				</map>

			</gender>

			<ignore xsi:type="string" column="ignoriert1"/>

			<ignore xsi:type="string" column="patfakt1"/>

		</idat>

		<!-- for MDAT in patient table, use the same patient-table also as

		a visit table (visit date needed). the patient id can be re-used

		as visit id, or a constant visit ID can be used. -->

	</patient-table>

	<!-- optional -->

	<visit-table>

		<source xsi:type="plain-file">

			<url>test-2-visits-duplicate.txt</url>

			<type>text/csv</type>

			<separator>\t</separator>

		</source>

		<idat>

			<patient-id column="patid"/>

			<visit-id column="fallnr"/>

			<start column="start" format="d.M.u[ H[:m[:s]]]" na="" truncate-to="year"/>

			<end column="end" format="d.M.u[ H[:m[:s]]]" na=""/>

		</idat>

		<mdat>

			<!-- in/out code -->

			<concept id="weight">

				<value column="gewicht" xsi:type="string"/>

				<!-- TODO: verify that na values are no allowed for concept start timestamps -->

				<start column="start" format="d.M.u[ H[:m[:s]]]" na=""/>

			</concept>

		</mdat>

	</visit-table>

	<wide-table>

		<source xsi:type="plain-file">

			<url>test-1-widetable.txt</url>

			<separator>\t</separator>

		</source>

		<idat>

			<patient-id column="patid"/>

			<visit-id column="fallnr"/>

		</idat>

		<mdat>

			<concept id="natrium">

				<value column="na" xsi:type="decimal"/>

				<start column="zeitpunkt" format="d.M.u[ H[:m[:s]]]"/>

				<unit constant-value="mmol/l" />

				<modifier id="other">

					<value xsi:type="string" na="" constant-value=""/>

				</modifier>

			</concept>

			<concept id="kalium">

				<value column="k" xsi:type="decimal"/>

				<start column="zeitpunkt" format="d.M.u[ H[:m[:s]]]"/>

			</concept>

			<concept id="enum1">

				<value column="en" na="" xsi:type="string">

					<map>

						<case value="1" set-value="" set-concept="enum1_1"/>

						<case value="2" set-value="" set-concept="enum1_2"/>

						<otherwise action="drop-fact" />

					</map>

				</value>

				<start column="zeitpunkt" format="d.M.u[ H[:m[:s]]]"/>

			</concept>

		</mdat>



	</wide-table>

	<eav-table>

		<source xsi:type="plain-file">

			<url>test-1-eavtable.txt</url>

			<separator>\t</separator>

		</source>

		<idat>

			<patient-id column="patid"/>

			<visit-id column="event"/>

			<ignore column="locat" xsi:type="string"/>

			<ignore column="user" xsi:type="string"/>

		</idat>

		<mdat>

			<concept column="param"/>

			<start column="start_ts" format="u-M-d"/>

			<end column="end" format="u-M-d" na="@"/>

			<!-- location -->



			<!-- if group is non-na, all following eav rows with the same group value

			are considered facts in a single group. A group must contain a primary

			concept and any number of modifiers. All rows within the group must have

			the same patient,encounter and start timestamp. -->

			<!-- <group column="group" na="@" /> -->



			<type column="type"/>

			<value column="value" na="@"/>

			<unit column="unit" na="@"/>

		</mdat>

		<!-- virtual value columns -->

		<virtual>

			<!-- for eav facts, the map always applies to the fact value -->

			<value column="f_eav_m" xsi:type="string">

				<map set-concept="f_eav_m_m" />

			</value>

			<value column="f_eav_x" na="" xsi:type="string">

				<map set-value="">

					<case value="1" set-concept="f_eav_x_1"/>

					<!-- action inplace is default -->

					<case value="0" set-concept="f_eav_x_0"/>

					<!-- action drop will not produce any fact -->

					<otherwise log-warning="Unexpected value" action="drop-fact" />

					<!-- action generate will produce a new concept with the given values -->

				</map>

			</value>

			<!-- ... more value elements -->

		</virtual>

	</eav-table>



</datasource>