Manipulating text with query expressions in Django

I was working with a table that had a similar (simplified) structure like this:

|               uuid               |         file_path         |
|----------------------------------|---------------------------|
| b8658dfc3e80446c92f7303edf31dcbd | media/private/file_1.pdf  |
| 3d750874a9df47388569a23c559a4561 | media/private/file_2.csv  |
| d177b7f7d8b046768ab65857451a0354 | media/private/file_3.txt  |
| df45742175d7451dad59761f15653d9d | media/private/image_1.png |
| a542966fc193470dab84351c15523042 | media/private/image_2.jpg |

Let’s say the above table is represented by the following Django model:

from django.db import models


class FileCabinet(models.Model):
    uuid = models.UUIDField(
        primary_key=True, default=uuid.uuid4, editable=False
    )
    file_path = models.FileField(upload_to="files/")

I needed to extract the file names with their extensions from the file_path column and create new paths by adding the prefix dir/ before each file name. This would involve stripping everything before the file name from a file path and adding the prefix, resulting in a list of new file paths like this: ['dir/file_1.pdf', ..., 'dir/image_2.jpg'].

Using Django ORM and some imperative Python code you could do the following:

...

# This will give you a queryset with the file paths.
# e.g. <QuerySet ['media/private/file_1.pdf', ... ]>
file_paths = FileCabinet.objects.values_list("file_path", flat=True)

# Now the file names can be collected in a list via a listcomp.
# This will return: ["dir/file_1.pdf", ..., "dir/image_2.jpg"]
file_paths_new = [
    f"dir/{file_path.split('/')[-1]}" for file_path in file_paths
]

...

Here, we use the FileCabinet model to make a query and obtain the file paths. We then use Python to split the file paths and extract the file names, and add the prefix dir/ to create the new paths. While this approach is relatively simple, it can be slow and resource-intensive if the size of the working dataset is large. This is because the entire working dataset is loaded into memory and the text manipulation is performed in Python.

To improve performance and efficiency, Django offers a declarative approach using expressions. These expressions allow you to offload operations like this to the database, which can be significantly faster and less resource-intensive than the imperative approach, especially for larger querysets. Here’s how you can achieve the same result in a declarative manner:

...

from django.db.models import F, Value
from django.db.models.functions import (
    Concat,
    Reverse,
    Right,
    StrIndex,
)


file_cabinet = polls_models.FileCabinet.objects.annotate(
    last_occur=StrIndex(Reverse(F("file_path")), Value("/")),
    file_name=Right(F("file_path"), F("last_occur") - 1),
    file_path_new=Concat(Value("dir/"), F("file_name")),
)

...

You can see the new file paths by inspecting the file_cabinet queryset as follows:

file_paths_new = file_cabinet.values_list("file_path_new", flat=True)

This will give you the following queryset:

<QuerySet
    ['dir/file_1.pdf',
    'dir/file_2.csv',
    'dir/file_3.txt',
    'dir/image_1.png',
    'dir/image_2.jpg']
>

Now, let’s step through the each of the ORM functionality that was levereged here:

The annotate function is being used to add additional information to each returned FileCabinet object. This function allows you to specify additional fields that should be calculated and included in the returned queryset. Inside the annotation method, we use F objects to reference a model field within the query. They can be used to refer to a field’s value in the context of an update or filter, rather than referring to the actual field itself.

Three fields are being added to the FileCabinet objects: last_occur, file_name, and file_path_new.

last_occur is being calculated by using the StrIndex function. This function takes two arguments: the string to search and the string to search for. In this case, the string being searched is the file_path field, but it has been passed through the Reverse function to reverse the string. This is done so that the StrIndex function starts searching from the end of the string, rather than the beginning. The second argument to StrIndex is the string to search for, which in this case is /. The StrIndex function returns the position of the first occurrence of the search string in the main string.

file_name is being calculated by using the Right function. This function takes two arguments: the string to extract from and the number of characters to extract. In this case, the string being extracted from is the file_path field, and the number of characters to extract is specified by the last_occur field. The last_occur field represents the position of the last occurrence of / in the file_path field, so extracting the characters from this position onwards gives us the file name with its extension. The - 1 at the end is used to remove the / character itself from the extracted string.

Finally, the file_path_new is constructed by using the Concat function. This function takes a variable number of arguments and concatenates them together into a single string. In this case, the dir/ prefix is being concatenated with file_name field.

Perfection!

Recent posts