Editing Command–query separation (section)

==Limitations==
CQS can introduce complexities for implementing [[Reentrant (subroutine)|reentrant]] and [[Multithreading (software)|multithreaded]] software correctly. This usually occurs when a non-thread-safe pattern is used to implement the command-query separation.

Here is a simple example that does not follow CQS, but is useful for multi-threaded software because it solves the complexity of locking for all other parts of the program, but by doing so it doesn't follow CQS because the function both mutates state and returns it:
<syntaxhighlight lang="java">
private int x;
public int incrementAndReturnX() {
    lock x;   // by some mechanism
    x = x + 1;
    int x_copy = x;
    unlock x; // by some mechanism
    return x_copy;
}
</syntaxhighlight>

Here is a CQS-compliant version. Note that it is safely usable only in single-threaded applications. In a multithreaded program, there is a race condition in the caller, between where <code>increment()</code> and <code>value()</code> would be called:
<syntaxhighlight lang="java">
private int x;
public int value() {
    return x;
}
void increment() {
    x = x + 1;
}
</syntaxhighlight>

Even in single-threaded programs, it is sometimes arguably significantly more convenient to have a method that is a combined query and command. [[Martin Fowler (software engineer)|Martin Fowler]] cites the <code>pop()</code> method of a [[Stack (abstract data type)|stack]] as an example.<ref>{{cite web|last=Fowler|first=Martin|title=CommandQuerySeparation|url=http://martinfowler.com/bliki/CommandQuerySeparation.html|access-date=5 December 2005}}</ref>