Speed up patent-client

Question

I am using patent-client in order to look for the patent number of 250000 doc-numbers. My original data look like this:

invention_title doc_number  date
0   Doughnut product with six appendages    29327507    2008
1   Doughnut product with six appendages and witho...   29327632    2008
2   Doughnut product with six appendages and witho...   29327637    2008
3   Meat piece  29298838    2007
4   Pet treat   29320494    2008
...

I am trying to use patent-client in order to retrieve the patent number of each observation starting from the doc-number (column doc-number) as follows:

# Import the model classes you need
from patent_client import Inpadoc, Assignment, USApplication

# Fetch US Applications
app={}
patent_x=[]
publn_nr=[]

for i in range(len(df_all)):
    try:
        app[i] = USApplication.objects.get(df_all['doc_number'][i])
    except:
        pass
    try:
        patent_x.append(app[i].patent_number)
    except:
        patent_x.append('')
    try:
        publn_nr.append(app[i].publication_number)
    except:
        publn_nr.append('')
        
        

df_all['patent_x']=patent_x
df_all['publn_nr']=publn_nr

However this code takes a huge amount of time (while if I look for single doc-number one at the time the procedure seems very fast). Why so? Is there a way I can improve the speed of the process?

Thank you