The Clients & Profits
database contains a set of data records corresponding to
one or more file formats (called "data files"),
along with a set of indexes for those records. Clients & Profits
stores its data files, field names, file connections (i.e.,
relations or joins) in a data dictionary inside the database.
Unlike relational databases like FoxPro and Access, Clients & Profits
keeps all of its data files, indexes, etc. in a single physical file on
the server's hard disk. The file has a name ending in .DF1 and looks like
this:
The Clients & Profits database stores records
for the 50+ data files used by the software to track clients,
jobs, costs, etc. A list of the data files used by Clients & Profits
can be found in the Database Utilities window (see right, or
choose Setup > Utilities > Database Utilities) or from the Data
Dictionary.
Each data file (e.g., JOBS, COSTS, etc.) creates
a slot in the database. Clients & Profits also reserves
some slots as system slots.
Database segments support large amounts
of data
A Clients & Profits database may comprise up to 15 separate physical
document files, called segments. Each individual segment has a maximum
size of 256 mb, and is identified with a special file extension (i.e.,
.DF2, .DF3, .DF4, etc.). Clients & Profits treats these segments as one
continuous database. Clients & Profits automatically expands the first
segment (i.e., .DF1) of the database as it needs more space by adding
unused disk blocks to it. This happens behind-the-scenes as users add new
records to the database. When the database nears 256 megabytes, an
additional segment (which will have the file extension .DF2) must be
created. These additional segments can't be created or expanded
automatically, so the Database Utilities window has a special "Change
Database Size" function that handles them.
The Database Utilities window in Clients & Profits offers system managers
a single console to monitor the database's growth, check its stability,
and change its size. It also includes tools for exporting and importing
data.
 Always monitor the database size. It's important to regularly check the
size of your Clients & Profits database, especially once it exceeds 200
megabytes. It's very easy to neglect the size until the 256 mb limit is
reached, which triggers the "data files full" error -- and can cause
damage to the database. It's easy to see the database size: it appears at
the bottom of the Introduce Yourself window every time someone starts
Clients & Profits. The Database Utilities window also has a function that
can graph the database's growth over time, helping you plan when to add
the next segment.
Slots, blocks, records, indexes, and pointers
The Clients & Profits database is composed of slots that contain the data
from the program's various files, like jobs, clients, costs, and G/L
entries. Each slot contains blocks that store bits of data from these
files. A single block stores 512 characters -- that's all. So a large
database can have tens of thousands of blocks, each storing 512 characters
(i.e., bytes).
There are five types of blocks in the C&P database:
Data
blocks
Index
blocks
File
header blocks
Free
blocks
The
"master" block
The block that stores the information for
a database record is a data block. Depending on the size
of an individual record, one data block may be large enough to
store several records from a data file (such as time entries,
which are small), or Clients & Profits might need several data
blocks to store just one record (such as purchase orders, which
are large).
Index blocks store index values. Every
data file has special index fields which are used for finding
records, such as the client number, job number, task, etc. Clients & Profits
indexes the data values in a tree structure. Every record added,
changed, or deleted rearranges the tree.
Free blocks are unused blocks in the
data file. Clients & Profits keeps track of the free blocks in
a data file in a free chain. Clients & Profits updates the free
chain whenever it deletes a record from a specific data file
by marking the block or partial block (depending on the size
of the record) that the deleted record occupied as being available
for new records.
Clients & Profits maintains a file header
block for each data file in the database. File header blocks
store information about the attributes of each field in a data
file. This lets C&P determine whether a particular data file
needs to be reorganized with each new update. The file header
block also maintains a pointer to the first block in the data
chain for that file, as well as a pointer to the first and
last block of all indexes for that file (and other information).
Block 0 is the master block for the
database, which contains critical information about the database.
It holds the total block count for its own segment of the database
and, if there is more than one segment in the database, also
holds information about the sizes of the other segments. Block
0 also stores a pointer in the first block in the data
set free chain, a directory of all files in the database,
and a point to each of the file header blocks.
The pointer is like a street address. It tells Clients & Profits where a
specific record (for example, a job ticket) is stored in the database.
When a user finds a job ticket, Clients & Profits requests the record from
file server using its address (i.e., by pointing to the block in the
database). Clients & Profits doesn't actually know where the data is
stored physically on the file server's hard disk. The only thing it knows
is the slot, block, and pointer -- the unique address for each record in
the database.
Managing multiple databases
A separate database file is created for each database you'll use, such as
the tutorial database, the sample database, and your shop's database. Each
database file works like any other application file; it is opened, moved,
and backed up like any word processor or spreadsheet document. If your
accounting department manages the books for several different companies,
you'll probably have several different database files on your server's
hard disk. Each database will have a unique file name, but will always end
in .DF1.
The Clients & Profits database can contain 16 segments, each containing up
to 256mb of clients, jobs, costs, and other data. Each database segment
has a special file name extension that identifies its order in the
database structure.
How Clients & Profits uses the Quick Check to look for trouble
The "quick check" function looks for early data corruption in the Clients
& Profits database. It runs automatically each time a manager-level user
clicks OK in the Introduce Yourself window.
The Quick Check examines the status of the database by reading only the
internal headers, indexes, and pointers (not the records themselves, to
save time). These records indicate corruption caused by either hardware,
network, or other system failures. It doesn't attempt to systematically
check the entire database for problems; that's the role of the Check Data
Files function in Database Utilities. For this reason, running a quick
check only takes a few seconds.
Because it is optimized for speed, the Quick Check doesn't find every kind
of database damage. However, it is very effective at finding data
corruption early, before it has a chance to affect the accuracy of the
data. If the Quick Check function discovers some corruption, it displays
what it found in a log window. If no one else is using the C&P database,
clicking the Repair button will attempt to fix the corruption immediately.
BE SURE TO BACK UP THE DATABASE BEFORE IT IS REPAIRED since any kind of
system crash during a repair function will make the database unusable.
Since the Quick Check runs every time a manager opens the database, it is
likely to find data corruption within minutes or hours of its occurrence.
This time frame is important because system managers can use it to track
down the cause of the corruption while the day's events are still fresh in
users' minds. In shops with many manager-level users, this means the quick
check is run dozens of times each day.
Next Page: Database
Size |
