Question

I’m creating some instances of the Item model shown below from a csv file. In the csv file only the name and the filename of the Item is specified. The image field is set by looking if the image exists, and if it does, set the image file to image:

I do:

item = Item()
item.name = csv_fields[0]
item.filename = csv_fields[1]
item.save()

...

f = open(settings.MEDIA_ROOT+"images/"+item.filename, 'r')
item.image = File(f)
item.save()

Doing so, the image is duplicated in settings.MEDIA_ROOT+"images/" as <original filename>_1. How can avoid that? i.e. how can I just set an existing file to an imagefield, without copying it?

class Item(models.Model):   

    name = models.CharField(max_length=50)
    filename = models.CharField(max_length=100,)
    image = models.ImageField(upload_to='images', blank=True, null=True)

Answer 1

This is all you need

item = Item()
item.name = csv_fields[0]
item.filename = csv_fields[1]
item.image = csv_fields[1]
item.save()

the image field takes a filename and the field definition’s upload_to, and converts that to an image resource. if you pass in an image resource, it thinks its making a new one.

SO this:

f = open(settings.MEDIA_ROOT+"images/"+item.filename, 'r')
item.image = File(f)

isn’t necessary at all

the only diff between a charfield and a image field is that a imagefield takes the filename and casts it to a PIL Image resource for further manipulation. its just a wrapper.