Table Of Contents

Data Files

File Names and Format

File Groups

Location (Sequence Number)

Client Archive Directory

Data Files

---

The WANDEST server unloads one group of data per upload or update request. The file format is identical to the format of the unload facility from Informix.

The server sends upload or update data to the client using the FTP protocol. The server puts files on the client in the directory specified by the WDCLIENT_OUTDIR parameter in the client Configuration file. The group and sequence numbers that are passed through the socket interface further determine the actual destination directories.

File Names and Format

Data is returned in ASCII format in files containing lines for each row of data. The column data for each row is separated by the delimiter character specified in the client Configuration file. Each row contains key values followed by the column values in the order specified in the request. Using the FTP command put, the files are transferred from the server to the client within this call.

A full upload generates one file for each table, <tablename>.unl, while an update generates two files for each table: one file for record insert or modify, <tablename>.new, and one file for record deletions, <tablename>.del.

An option exists for transferring all upload and update tables into one file. Set this option with the WDCLIENT_ONEFTP parameter in the client Configuration file. With this parameter, the data files for all tables are zipped together, and the server uses only one thread to FTP the zipped file to the client.

Table 4-1 lists the file names for an upload and update.

Table 4-1 Upload and Update File Names

Function	File Name
File Upload One File Per Table Option	If DBPopulate = Y, tablename.yyyy.mo.dd.hr.mm.ss If DBPopulate = N, tablename.yyyy.mo.dd.hr.mm.ss.Z
File Upload All Tables in One File Option After tables are untarred	fullupload.<timestamp>.tar If DBPopulate = Y, tablename.unl If DBPopulate = N, tablename.unl.Z
Incremental Update Two Files Per Table Option Insert/Update Delete	If DBPopulate = Y new.tablename.yyyy.mo.dd.hr.mm.ss delete.tablename.yyyy.dd.hh.mm.ss If DBPopulate = N: new.tablename.yyyy.mo.dd.hr.mm.ss.Z delete.tablename.yyyy.dd.hh.mm.ss.Z
Incremental Update All Tables in One File Option After tables are untarred	inc_upload.<timestamp>.tar If DBPopulate = Y tablename.new tablename.del If DBPopulate = N tablename.new.Z tablename.del.Z

---

Note The .Z extension indicates a compressed file format. The .Z extension is not used on tar files.

---

Table 4-2 lists the file naming conventions.

Table 4-2 File Naming Conventions

Character	Definition
yyyy	Year of the start of process in GMT time
mo	Month
hr	Hour
mm	Minute
ss	Seconds

File Groups

Program the desired tables into file groups, as the WANDEST client requests one group of uploads or updates per request. Groups are programmed in the Control file. For information about the Control file, see "WANDEST Client".

For incremental updates, certain tables might be uploaded more frequently than others. Changes occur more frequently in these tables; therefore, the corresponding server tracking table contains many entries. Groups of tables can be formed, enabling each group to have its own incremental update interval.

For example, during provisioning, many entries are added to the CWM user_connection table. Program this table and all segment tables (connection, atm_connection) into a group called Group1. Program other less-frequently updated tables, such as packet_line and physical_line table into Group2. Specify a shorter update interval for Group1 to keep the user_connection data more consistent with the CWM database.

---

Note The grouping concept applies only to the WANDEST client, not to the server.

---

Every table name is followed by the group ID to which that table belongs.

---

Note If the group ID in the Control file is blank, the full upload and incremental upload does not work.

---

One group can have several tables, or all tables can belong to one group. The group name is followed by the column names to be extracted. Specify the name and data type of each column. Data types are provided in Table 3-1. When no data is needed from a table, do not include a table name.

For example, the database tables are separated into 10 groups. To perform a full upload of the entire database, request a full update for each of the 10 groups. Similarly, for an incremental update of the entire database, request an incremental update for each of the 10 groups.

Location (Sequence Number)

The server places the data files in a parent directory specified by the WDCLIENT_OUTDIR parameter in the client Configuration file. Further directories are specified by the group and sequence numbers that are the parameters for a full upload or an incremental update.

For example, /usr/users/wandest/client/out is the output directory specified in the WDCLIENT_OUTDIR parameter. The destination directory for the full upload for Group1, sequence4, is placed in /usr/users/wandest/client/out/Group1/4.

The WDCLIENT_MAXDIRSEQ parameter allows a range of sequences, so changes occurring to the same groups can be placed in different directories for analysis.

Client Archive Directory

If the WANDEST client issues a command that has been previously issued, for example a full upload for Group5, sequence6, the client checks the destination directory for any files that are present. Previously issued commands result in files being deposited in the destination directory. Thus, the client moves the older files in the destination directory to the archive directory to make room for the results of the current command.

The WANDEST client moves the older files to the archive directory specified by the WDCLIENT_ARCHIVEDIR parameter in the client Configuration file. The default directory is /usr/users/wandest/client/data/save. The client appending the files in the archive directory.

---

Note Users are responsible for cleaning up the data files in the archive directory.

---