-
Notifications
You must be signed in to change notification settings - Fork 85
Description
Description:
I am using node-odbc to connect to a Sybase ASE database. I am encountering an issue where string data containing special characters or specific encodings is returned with substitute characters (\x1A) or corrupted text.
I have verified that the exact same query and database setup works correctly using msnodesqlv8 package. Since msnodesqlv8 retrieves the correct Unicode/string data within the same Node.js environment, this suggests the issue is specific to how node-odbc is binding the columns (likely defaulting to SQL_C_CHAR when it needs SQL_C_WCHAR or vice versa) or handling the charset conversion from the ODBC driver.
Environment:
OS: Windows 10
ODBC Driver: Adaptive Server Enterprise / FreeTDS
Database: Sybase ASE
Reproduction Steps:
Connect to Sybase ASE using node-odbc.
Select data from a column containing non-ASCII characters.
Observe artifacts (\x1A) in the output.
Run the same query using msnodesqlv8 and observe correct output.
Comparison Code:
const odbc = require('node-odbc');
const sql = require('msnodesqlv8');
const query = "SELECT test_column FROM MyTable WHERE id = 1";
// 1. Test with msnodesqlv8 (Works)
const connectionStringMS = "Driver={...};Server=..."; // Your working connection string
sql.query(connectionStringMS, query, (err, rows) => {
if (err) console.error(err);
console.log('msnodesqlv8 result:', rows);
// Output: Correct String (e.g., "München")
});
// 2. Test with node-odbc (Fails)
async function runOdbc() {
const connection = await odbc.connect('DSN=MySybaseDSN;UID=...;PWD=...');
const result = await connection.query(query);
console.log('node-odbc result:', result);
// Output: Corrupted String (e.g., "München" becomes "M\x1A - \x1A")
}
runOdbc();Expected Behavior:
node-odbc should return the string correctly, matching the behavior of msnodesqlv8.
Actual Behavior:
node-odbc returns data containing \x1A (SUB) characters. This implies a mismatch in the column binding type (ODBC C-Type) requested by node-odbc versus what the driver provides.
Possible Cause:
It appears node-odbc is not correctly inferring the need for Wide Character binding (SQL_C_WCHAR) for this driver, whereas msnodesqlv8 (via OLE DB) handles the buffer width correctly.