Prev: deriving a class with a subset of base class methods?
Next: template class specialization trouble
From: sjefet on 7 Apr 2008 01:13
Hi,

I am new to boost, and to templates (feeling very comfortable with C).
I am trying to use boost hash tables where the keys are char *, like:

unordered_map<const char *, CMyclass *> myclass_map_t ;

1. Couldnt figure so far, how do i set a hash function that works on
the string (and not on the pointer) ?
2. Whenever i get a new key i want to check first if it is already in
the hash table, if yes - get its value, and if not have the
opportunity to create a new value for it, but be able to do it with
one operation. Is there a way to achieve that ?

TIA,

jehs.

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

From: Daniel Krügler on 7 Apr 2008 07:37
On 7 Apr., 18:13, sje...(a)gmail.com wrote:
> I am new to boost, and to templates (feeling very comfortable with C).
> I am trying to use boost hash tables where the keys are char *, like:
>
> unordered_map<const char *, CMyclass *> myclass_map_t ;

Note: I have no experience with this new facility of the
boost trunk, but it seems that it reflects the current
specification of upcoming C++0x std::unordered_map (given
current compiler technologies, i.e. modulo rvalue references
where not available).

> 1. Couldnt figure so far, how do i set a hash function that works on
> the string (and not on the pointer) ?

IMO hash functions on string-like entities belong to
the most popular use-cases of hashers. A simple way would
be to have a look at boost's implementation for std::string,
which effectively calls boost::hash_range. So you could provide
something like

#include <string.h>

struct CharArrayHashEqual : std::unary_function<const char*,
std::size_t>
{
std::size_t operator()(const char* s) const {
return boost::hash_range(s, strlen(s));
}

bool operator(const char* s1, const char* s2) const {
return strcmp(s1, s2) == 0;
}
};

and use type CharArrayHash as the third (Hash) *and*
forth (EqualTo) template argument of boost::unordered_map.

Also very popular are the FVN algorithm,

http://www.isthe.com/chongo/tech/comp/fnv/

or the Jenkins hash,

http://burtleburtle.net/bob/hash/doobs.html

(instead of the "boost::hash_range(s, strlen(s))" part)

> 2. Whenever i get a new key i want to check first if it is already in
> the hash table, if yes - get its value, and if not have the
> opportunity to create a new value for it, but be able to do it with
> one operation. Is there a way to achieve that ?

You will probably write a helper function which does
this, which e.g. uses the member function insert that
returns a pair<iterator, bool>. Depending on the bool
part of the return value you can modify the value
accessible via the iterator result value or not.

HTH & Greetings from Bremen,

Daniel Kr�gler


--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

From: Daniel Krügler on 7 Apr 2008 23:06
On 8 Apr., 00:37, Daniel Kr�gler <daniel.krueg...(a)googlemail.com>
wrote:
> #include <string.h>
>
> struct CharArrayHashEqual : std::unary_function<const char*,
> std::size_t>
> {
> std::size_t operator()(const char* s) const {
> return boost::hash_range(s, strlen(s));

This should be

return boost::hash_range(s, s + strlen(s));

of course.

Sorry for the noise,

Daniel


--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

From: Mathias Gaunard on 7 Apr 2008 23:06
On Apr 8, 12:37 am, Daniel Kr�gler <daniel.krueg...(a)googlemail.com>
wrote:

> struct CharArrayHashEqual : std::unary_function<const char*,
> std::size_t>
> {
> [...]
> bool operator(const char* s1, const char* s2) const

What is that supposed to declare?



--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

From: Daniel Krügler on 7 Apr 2008 23:18
On 8 Apr., 00:37, Daniel Kr�gler <daniel.krueg...(a)googlemail.com>
wrote:
> On 7 Apr., 18:13, sje...(a)gmail.com wrote:
>
> > I am new to boost, and to templates (feeling very comfortable with C).
> > I am trying to use boost hash tables where the keys are char *, like:
>
> > unordered_map<const char *, CMyclass *> myclass_map_t ;
>
> Note: I have no experience with this new facility of the
> boost trunk, but it seems that it reflects the current
> specification of upcoming C++0x std::unordered_map (given
> current compiler technologies, i.e. modulo rvalue references
> where not available).
>
> > 1. Couldnt figure so far, how do i set a hash function that works on
> > the string (and not on the pointer) ?
>
> IMO hash functions on string-like entities belong to
> the most popular use-cases of hashers. A simple way would
> be to have a look at boost's implementation for std::string,
> which effectively calls boost::hash_range. So you could provide
> something like
>
> #include <string.h>
>
> struct CharArrayHashEqual : std::unary_function<const char*,
> std::size_t>
> {
> std::size_t operator()(const char* s) const {
> return boost::hash_range(s, strlen(s));
> }
>
> bool operator(const char* s1, const char* s2) const {
> return strcmp(s1, s2) == 0;
> }

Hastiness doesn't pay well:

std::size_t operator()(const char* s) const {
return boost::hash_range(s, s + strlen(s));
}

bool operator()(const char* s1, const char* s2) const {
return strcmp(s1, s2) == 0;
}

Greetings from Bremen,

Daniel Kr�gler



--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

 |  Next  |  Last
Pages: 1 2
Prev: deriving a class with a subset of base class methods?
Next: template class specialization trouble