I’m interested in conditioning on the values of only a subset of variables in a random variable array. The example for condition in the docs goes something like this:

```
from jax import random
import numpyro
from numpyro.handlers import condition, seed, substitute, trace
import numpyro.distributions as dist
def model():
numpyro.sample('a', dist.Normal(0., 1.))
model = seed(model, random.PRNGKey(0))
exec_trace = trace(condition(model, {'a': -1})).get_trace()
exec_trace['a']['value']
```

```
-1
```

I’m interested in the following generalization to multiple random variables:

```
from jax import random
import numpyro
from numpyro.handlers import condition, seed, substitute, trace
import numpyro.distributions as dist
def model():
numpyro.sample('a', dist.Normal(0., 1.).expand([2]))
model = seed(model, random.PRNGKey(0))
exec_trace = trace(substitute(model, {'a': [-1, np.nan]})).get_trace()
exec_trace['a']['value']
```

But the output that I get is:

```
[-1, nan]
```

Instead of, for example:

```
[-1, 0.284742]
```

Does anyone know how to condition on the values of just some of the variables in an array? Just inserting `np.nan`

doesn’t work and I couldn’t find any examples in the documentation or on the past posts in the forum. Any and all help is very much appreciated.