Extracting information from JSON file using python

I have a very large dataset (JSON files) which consists of images, questions, and answers. I want to select the images and questions which have single word answers. I want the result form V2.0 VQA dataset, You can enter this website to have a clear idea about the dataset: [login to view URL] I have use the codes here (VQA API) to get data from the dataset: [login to view URL] You have to download questions and answers from the following link: (Very large files I cannot attach them) [login to view URL] I want Balanced Real Images VQA Annotations: v2.0 for both train and val VQA Input Questions: v2.0 for both train and val I have this idea but I could not apply it: In order to retrieve the list of all question ids that have single word answer, you need to check for the length of the most common answer (denoted by the key "multiple_choice_answer") for each question. So essentially you need to add the following line -- anns = anns if len(ansLengths) == 0 else [ann for ann in anns if len(ann['multiple_choice_answer'].split()) in ansLengths] where ansLengths is expected to be a list of numbers specifying the desired lengths. The line is needed to be added in the API: [login to view URL] Alternatively, you could skip using the API; just download the data and loop through all the question ids and count those whose "multiple_choice_answer"'s are one word in length. Thanks