I m writing a small utility that captures and logs SQL statements, but will have to remove sensitive data from the Query text and replace with with some dummy text (i.e:XXXXX).
What is a good way to parse the SQL query in java and replace parameters value?
for example:
replace
SELECT NAME, ADDRESS, .... FROM USER WHERE SSN IN ('11111111111111', '22222222222222');
with
SELECT NAME, ADDRESS, .... FROM USER WHERE SSN IN (?, ?);
Using JSQLParser (V0.8.9) this is a solution for your problem:
String sql ="SELECT NAME, ADDRESS, COL1 FROM USER WHERE SSN IN ('11111111111111', '22222222222222');";
Select select = (Select) CCJSqlParserUtil.parse(sql);
//Start of value modification
StringBuilder buffer = new StringBuilder();
ExpressionDeParser expressionDeParser = new ExpressionDeParser() {
@Override
public void visit(StringValue stringValue) {
this.getBuffer().append("XXXX");
}
};
SelectDeParser deparser = new SelectDeParser(expressionDeParser,buffer );
expressionDeParser.setSelectVisitor(deparser);
expressionDeParser.setBuffer(buffer);
select.getSelectBody().accept(deparser);
//End of value modification
System.out.println(buffer.toString());
//Result is: SELECT NAME, ADDRESS, COL1 FROM USER WHERE SSN IN (XXXX, XXXX)
This replaces all found String values within your SQL. To replace other types of data e.g. Long values, override the corresponding visit method in ExpressionDeParser.
Don't use regexp in this case. It turns out quickly to be hard maintainable.
The correct answer depends on how much you want to replace. Something like:
[0-9]{3}-?[0-9]{2}-?[0-9]{4}
will replace social security numbers pretty well. I always take regex code to
to tweak it and work out bugs.
If you need to replace tons of sensitive information though, and if there are a lot of cases, definitely start looking into using a parser to parse the SQL query string. (such as jsqlparser, as Anirudh recommended.)
String sqlDebit = select * from table where and billing_cycle_start_date between :startDate and :endDate
java:
sqlDebit= sqlDebit.replaceAll(":startDate ", ""+startDate).replaceAll(":endDate", ""+endDate);
With prepare statement you can replace "?" in your query string with your value. Use number to specify which "?" you are referring too. They go by order from right to left.
For example: "SELECT LastName, FirstName FROM Person.Contact WHERE LastName = ? and FirstName = ?"
pstmt.setString(1, "LastNameValue"); pstmt.setString(2, "FirstNameValue"); see full example below:
public static void executeStatement(Connection con) {
try(PreparedStatement pstmt = con.prepareStatement("SELECT LastName, FirstName FROM Person.Contact WHERE LastName = ?");) {
pstmt.setString(1, "Smith");
ResultSet rs = pstmt.executeQuery();
while (rs.next()) {
System.out.println(rs.getString("LastName") + ", " + rs.getString("FirstName"));
}
}
// Handle any errors that may have occurred.
catch (SQLException e) {
e.printStackTrace();
}
}
