# dswap
> Interchange two double-precision floating-point vectors.
## Usage
```javascript
var dswap = require( '@stdlib/blas/dswap' );
```
#### dswap( x, y\[, dim] )
Interchanges two double-precision floating-point vectors `x` and `y`.
```javascript
var Float64Array = require( '@stdlib/array/float64' );
var array = require( '@stdlib/ndarray/array' );
var x = array( new Float64Array( [ 4.0, 2.0, -3.0, 5.0, -1.0 ] ) );
var y = array( new Float64Array( [ 2.0, 6.0, -1.0, -4.0, 8.0 ] ) );
dswap( x, y );
var xbuf = x.data;
// returns [ 2.0, 6.0, -1.0, -4.0, 8.0 ]
var ybuf = y.data;
// returns [ 4.0, 2.0, -3.0, 5.0, -1.0 ]
```
The function has the following parameters:
- **x**: a non-zero-dimensional [`ndarray`][@stdlib/ndarray/ctor] whose underlying data type is `float64`. Must have the same shape as `y`.
- **y**: a non-zero-dimensional [`ndarray`][@stdlib/ndarray/ctor] whose underlying data type is `float64`. Must have the same shape as `x`.
- **dim**: dimension along which to interchange vectors. Must be a negative integer. Negative indices are resolved relative to the last array dimension, with the last dimension corresponding to `-1`. Default: `-1`.
For multi-dimensional input [`ndarrays`][@stdlib/ndarray/ctor], the function performs batched computation, such that the function interchanges each pair of vectors in `x` and `y` according to the specified dimension index.
```javascript
var Float64Array = require( '@stdlib/array/float64' );
var array = require( '@stdlib/ndarray/array' );
var opts = {
'shape': [ 2, 3 ]
};
var x = array( new Float64Array( [ 4.0, 2.0, -3.0, 5.0, -1.0, 3.0 ] ), opts );
var y = array( new Float64Array( [ 2.0, 6.0, -1.0, -4.0, 8.0, 2.0 ] ), opts );
var v1 = x.get( 0, 0 );
// returns 4.0
var v2 = y.get( 0, 0 );
// returns 2.0
dswap( x, y );
v1 = x.get( 0, 0 );
// returns 2.0
v2 = y.get( 0, 0 );
// returns 4.0
```
## Notes
- Both input [`ndarrays`][@stdlib/ndarray/ctor] must have the same shape.
- Negative indices are resolved relative to the last [`ndarray`][@stdlib/ndarray/ctor] dimension, with the last dimension corresponding to `-1`.
- For multi-dimensional [`ndarrays`][@stdlib/ndarray/ctor], batched computation effectively means swapping all of `x` with all of `y`; however, the choice of `dim` will significantly affect performance. For best performance, specify a `dim` which best aligns with the [memory layout][@stdlib/ndarray/orders] of provided [`ndarrays`][@stdlib/ndarray/ctor].
- `dswap()` provides a higher-level interface to the [BLAS][blas] level 1 function [`dswap`][@stdlib/blas/base/dswap].
## Examples
```javascript
var discreteUniform = require( '@stdlib/random/array/discrete-uniform' );
var ndarray2array = require( '@stdlib/ndarray/to-array' );
var array = require( '@stdlib/ndarray/array' );
var dswap = require( '@stdlib/blas/dswap' );
var opts = {
'dtype': 'float64'
};
var x = array( discreteUniform( 10, 0, 100, opts ), {
'shape': [ 5, 2 ]
});
console.log( ndarray2array( x ) );
var y = array( discreteUniform( 10, 0, 10, opts ), {
'shape': x.shape
});
console.log( ndarray2array( y ) );
dswap( x, y );
console.log( ndarray2array( x ) );
console.log( ndarray2array( y ) );
```
[blas]: http://www.netlib.org/blas
[@stdlib/ndarray/ctor]: https://www.npmjs.com/package/@stdlib/ndarray-ctor
[@stdlib/ndarray/orders]: https://www.npmjs.com/package/@stdlib/ndarray-orders
[@stdlib/blas/base/dswap]: https://github.com/stdlib-js/blas/tree/main/base/dswap
[@stdlib/blas/gswap]: https://github.com/stdlib-js/blas/tree/main/gswap
[@stdlib/blas/sswap]: https://github.com/stdlib-js/blas/tree/main/sswap