I've noticed strange behavior when using a table comparison element to sync tables with Data Services 4.2 SP8 P2. Here's a demo scenario to explain (Oracle):

First create the required source and target tables

CREATE TABLE TEST_SOURCE
(
  ID      NUMBER,
  CODE_A  VARCHAR2(50 BYTE),
  CODE_B  VARCHAR2(50 BYTE)
);
CREATE UNIQUE INDEX TEST_SOURCE_PK ON TEST_SOURCE(ID);
ALTER TABLE TEST_SOURCE ADD (
  CONSTRAINT TEST_SOURCE_PK
  PRIMARY KEY
  (ID)
  USING INDEX TEST_SOURCE_PK
  ENABLE VALIDATE);

CREATE TABLE TEST_TARGET
(
  ID      NUMBER,
  CODE_A  VARCHAR2(50 BYTE),
  CODE_B  VARCHAR2(50 BYTE)
);
CREATE UNIQUE INDEX TEST_TARGET_PK ON TEST_TARGET(ID);
ALTER TABLE TEST_TARGET ADD (
  CONSTRAINT TEST_TARGET_PK
  PRIMARY KEY
  (ID)
  USING INDEX TEST_TARGET_PK
  ENABLE VALIDATE);

Next create a new data flow in DS "DF_TEST_SYNC" to perform the necessary inserts, updates and deletes to sync target with source:

Q_sort: order data by the "ID" column.
Table_Comparison: "ID" set as pk column for the comparison

Default target table options are kept.

Test no.1

Prepare the source table by inserting a million rows:

truncate table TEST_SOURCE;
INSERT INTO TEST_SOURCE
SELECT LEVEL, 'CODE_A_' || LEVEL, 'CODE_B_' || LEVEL 
FROM DUAL 
CONNECT BY LEVEL <= 1000000;
COMMIT;
-- make sure the target table is empty
truncate table TEST_TARGET;

Sync both tables by running the data flow
Number of rows in target: 1.000.000 => OK
Delete the FIRST 50.000 ID's from the source table:

delete from TEST_SOURCE where ID <= 50000;
commit; 

Run the data flow again => target now has 950.000 rows => OK

Test no.2

Reset source and target table:

truncate table TEST_SOURCE;
INSERT INTO TEST_SOURCE
SELECT LEVEL, 'CODE_A_' || LEVEL, 'CODE_B_' || LEVEL 
FROM DUAL 
CONNECT BY LEVEL <= 1000000;
COMMIT;
-- make sure the target table is empty
truncate table TEST_TARGET;

Sync both tables by running the data flow
Number of rows in target: 1.000.000 => OK
Delete the LAST 50.000 ID's from the source table:

delete from TEST_SOURCE where ID > 950000;
commit; 

Run the data flow again => target now has 999.473 rows => NOT OK, DS has detected rows to delete, but only a fraction of the 50.000 rows that need to be removed.

Run the data flow again => target now has 998.869 rows. This continues until eventually all 50.000 rows are deleted.

I've seen a similar issue with DS 4.1: https://archive.sap.com/discussions/thread/3948660

Has anyone else encountered this problem with Data Services?