# Using the ODBC jEDI

Read Time: 5 minute(s)

# Using the ODBC jEDI with an existing table

Creating an ODBC type file from an existing table can be as simple as:


This creates a stub that maps to the tablename parameter, and generates a dictionary for each table column so they're ready for querying with jQL.

# Note #1

A comma separated map/definition is generated in lieu of an existing csv. If you only want to map to a subset of the table, you can specify a column list and an optional primary key (if none of the columns are primary keys).

# Example #1

You might have a table of ORDERS containing an auto sequenced key, an ORDER_NBR column, and some other columns you wish to combine into the ODBC jEDI view of the table:


The example above will create a stub pointing to the ORDERS table, but will treat the ORDER_NBRcolumn as the key to the jBASE dynamic record. The PRODUCT, PRICE, and DATE_ORDERED columns will be mapped to attributes 1, 2, 3 and 4 (datetime fields are by default split into two fields, therefore the date portion maps to attribute 3 and the time portion to attribute 4).

# Steps to create an ODBC jEDI stub using a CSV

To generate a valid csv mapping file, and map it into your desired *RDBMS,*do the following in jBASE:

  1. CREATE-FILE DICT filename 1

  2. Create a dictionary for each table column you wish to map. The ID for each dictionary should be a valid column name. The attribute can be whatever you want, however the primary key column must be 0. The dictionary type should be A if using the classic PICK style, or D for the Prime/Universe style. The heading, justification, and width attributes are not used by the jEDI interface, only for running jBASE queries.

# Note #2

You can create other dictionaries for reporting purposes but you should use S (for classic PICK style) or V / I for Prime/Universe style.

  1. jCreateCSV {-O} -Dtype filename tablename.csv

The above command reads in all relevant dictionaries for filename, and generates a csv in the CSVdir parameter value defined in the jEDIdrivers.inifile.

# Note #3

It's recommended the csv files are named according to their intended table name (which could be the same as the filename). However, this is not an absolute requirement. Additionally, you can modify the resulting csv if desired using any jBASE or operating system text editor.

4. CREATE-FILE DATA filename TYPE=ODBC {additional qualifiers}

# Note #4

The above example assumes the csv is filename.csv The EXISTING argument is optional (the default value is set to YES, meaning the table exists; NO means the table does not exist, and will be created).

# ODBC CREATE-FILE Arguments and Information

Additional arguments you can use in the CREATE-FILE include:

CSV=csvname The csv option is required if the csv is not filename.csv
Table=tablename The tablename if not the same as filename
UniqueKey=YES|NO Defaults to YES. NO is used if the key column (i.e. attribute 0) is not a unique value per record/row
Existing=YES|NO Specifies whether the table already exists, or should be created
WRITEOPTS=[I,U,D] Used to restrict updates to a given table. The argument can use any combination of the following values:
For example, WRITEOPTS=U would only allow updates, WRITEOPTS=ID would allow inserts and deletes, and WRITEOPTS= would not allow any updates
Connect=dsn_label Specifies which database connection information string to use from the jEDIdrivers.ini file (default is used if this argument is not specified)

# Note #5

The attribute mapping from SQL column to multi-value record is arbitrary as there are no duplicates

# CSV/Schema Cleansing

When creating a csv definition from an existing jBASE file, the dictionaries used to generate the csv may cause a table column to not be 100% compliant with the data in the original dynamic-array record (e.g. typically resulting in a {VAR}CHAR type of an insufficient size).

The jCheckSchema command can be used to verify and even update the csv so that all the original records can be written to the RDBMS.

jCheckSchema original_jBASE_file {schema_path} schema_def {ID} {options}

If ID is specified, only that record (keyed by ID) is checked. Otherwise, all records (or those in the current active select list) are checked.

jCheckSchema options:

-alist Ignore "undefined" for the attributes specified in list(comma separated)
-c Check data integrity
-i Interactive: prompts for changes
-q Quiet: no progress displayed
-rn Round column lengths up to nearest factor of n
-w Disable control character warnings
-R Report only

# Example #2

jCheckSchema CUSTOMER CUSTOMER.csv -r5

This will read all records in the CUSTOMER file and check that the types and lengths of each attribute are compliant with the csv definition.

# Note #6

jCheckSchema will detect multi{sub}-value fields that have not been defined, however, it won't define them for you. You will need to amend the csv and specify multi{sub}-value association groups or positional multi-value options (if they are not repeating multi{sub}-values).


The jBC (BASIC) language has an IOCTL function which has the syntax of:

result_code= IOCTL(file_var, IOCTL_option, in_out_arg)

The ODBC jEDI has the following non-standard IOCTL options:

68 Over-riding SQL SELECT statement (in_out_arg) for the next SELECT file_var statement (note: the result set must match the type and number of columns expected for the primary key denoted by the def in the stub.
98 Specify the in_out_arg number of writes before a commit will be performed (useful for improved performance when performing a high volume of insert/update/delete operations).
99 Issue a commit.
7000 Execute the in_out_arg as an SQL statement on the target RDBMS (typically a singleton query-statement)

Introduction to the ODBC jEDI

Installing the ODBC jEDI on Linux/UNIX

Installing the ODBC jEDI on Windows

Structure of ODBC jEDI Initialization Files

Structure of ODBC jEDI Mapping Files

Configuring the ODBC jEDI on Linux/UNIX

Configuring the ODBC jEDI on Windows

Back to ODBC