Programming

Hello,

Could you please tell me if there are any books/tutorial on
gdbm programming with C for beginners.

So far, I've found few tutorials and one of those:
http://www.network-theory.co.uk/docs...cintro_22.html
is easy to understand but very short in treatment as it dealt with
three api(s) and others seem difficult so it would better to follow
books/detail tutorial.


Thanks.

It is good to know relational databases, generally, and SQL, and simpler C/C++/JAVA APIs, before diving into the combination. The APIs are more orderly in C++/JAVA, while C APIs run on assumptions and man pages. Starting with a simple API to learn C/C++/JAVA is a good start, before going into complex APIs.

Generally, gdbm allows you to create container objects with hash based search. Ther are many different container types. They all act much like a RDBMS table with one index only. You can index data a second way by creating a second, indexing container to search using a different key. Data storage can be by pointer outside the containers, so two containers can point to the same row or tuple of data. However, you need reverse pointers in the data so a delete on one key can delete in the other container, too.

Besides hash, which is great for fast random lookup but requires some configuration for size, there is linear-hash, tree, unsorted, array, linked-list (one way or forward-backward) containers, each with different advantages. Trees are always sorted and tolerate all amounts of data well without configuration. Unsorted linked list is very fast to insert and slow to query, update, delete specific item. Hash has a patterened disorder when examining all items.

Thank you. I'm looking for something like storing and retrieving few names and telephone numbers through c code.

I'm planning to use sqlite but like to test dbm/gdmb first as it doesn't use any dedicated shell or clent/server kind of thing, so a small database can be created without much overhead. I tried nosql mongodb but, as it uses javascript for its query language and I
don't know javascript so couldn't proceed to nosql.

The erliest and most basic are the tsearch(), bsearch(), lsearch(), hsearch(), qsort() family:
Man Page for tsearch (linux Section 3) - The UNIX and Linux Forums

C++ Standard Template libs has many very basic containers (no hash):
STL Containers - C++ Reference
map - C++ Reference

To get the hash, many of my employers have had roguewave h++ libraries: RWTPtrHashTable<T>

Don't overlook the fine container objects in JAVA, like tree map:
TreeMap (Java Platform SE 6)

Google reveals a myriad of open source containers/lists/maps/trees/tables. Some RDBMS implement a multi-column index as containers of containers of containers, so each column can be looked up even if others are not there (so the whole container is iterated). Many good wiki on the high level concepts of hash and such, very good to know first, so API controls make sense.

Finally, awk and bash have associative arrays or lookup vectors, which are hash maps (learn the lingo or get out). Just be careful you are not in the simple "array addressed by integer" tutorial! I am sure PERL has all this, too, so read the wiki and pick your poison. Do you grasp the array, linked list (sorted and unsorted), tree and hash basic concepts?

Thanks for your detailed and meticulous explanation.
I've found two tutorial sets that can be of good help but those are on
dbm rather than gdbm so little bit of R&D needs to be done.

Berkeley DB Reference Guide: Adding elements to a database

The content the following link points has got very striking similarities with
a book, Beginning Linux Programming by Mathew and I got confused as I failed to understand
who copied whom. Did Professor Fuller copy the content of the book or Neil Mathew
et al copy the content of the tutorial note for his book?

http://snap.nlc.dcccd.edu/learn/full...ap7/chap7.html

OO programming says lists should be in container objects. I saw a huge case statement several places today, and each should have been something like a containater lookup. It is less code and faster! Done right, the compiler can do the searches at compile time.

BDB is a modest RDBMS. but gdbm is GNU dbm is more at the container level.

If you can manage persistent objects, perhaps storing them in mmap()'s of flat files, then your containers can have the behaviors of an RDBMS.

What kind of container can be searched at compile time?