How do I delete all records where my date_ymd column in Azure Search Index is equal to a specific date?

Hi, I have an issue where I'm collecting a lot of data on a scheduled script that will store the data in a the Azure Search Index as a vector db and I then use RAG to collect data from this based on a prompt sent by a user in a chatbot and the AI's response is based on the top N documents from the storage.

My first question is if there's a way to workaround just using the top N documents from Azure Search Index. I get the top 50 documents for a users prompt and feed it to my chatbot's system prompt. Is there anyway to link my streamlit chatbot to the Azure Search Index directly without feeding only N documents.

Secondly and more importantly. I need the code to delete all documents for specific date_ymd values. Keep in mind all my columns and keys are string type and not date type. So Can you help with creating the script to delete values which have a certain string date.

This is my code for uploading documents:

def chunk_data(data, chunk_size):
    for i in range(0, len(data), chunk_size):
        yield data[i:i + chunk_size]


def upload_documents_to_search_client(df, embeddings_dict, chunk_size=32000):
    """Uploads documents with embeddings to the search client in chunks."""
    data = [
        {
            "@search.action": "mergeOrUpload",
            "hardware_id": str(row["hardware_id"]),
            "text_feedback": str(row["text_feedback"]) if "text_feedback" in row else "",
            "uninstall_text_feedback": str(row["uninstall_text_feedback"]) if "uninstall_text_feedback" in row else "",
            "os": str(row["os"]) if "os" in row else "",
            "date_ymd": str(row["date_ymd"]) if "date_ymd" in row else "",
            "Feature_Category": str(row["Feature_Category"]) if "Feature_Category" in row else "",
            "Sentiment": str(row["Sentiment"]) if "Sentiment" in row else "",
            "country": str(row["country"]) if "country" in row else "",
            "aiid": str(map_aiid_to_label(row["aiid"])) if "aiid" in row else "",
            "version_app": str(row["version_app"]) if "version_app" in row else "",
            "os_version": str(row["version"]) if "version" in row else "",
            "architecture": str(row["architecture"]) if "architecture" in row else "",
            "score": str(row["score"]) if "score" in row else "",
            "region": str(row["region"]) if "region" in row else "",
            "city": str(row["city"]) if "city" in row else "",
            "vector_text_feedback": next(
                (item["embeddings"].get("vector_text_feedback", []) for item in embeddings_dict if item["hardware_id"] == str(row["hardware_id"])),
                []
            ),
            "vector_uninstall_feedback": next(
                (item["embeddings"].get("vector_uninstall_feedback", []) for item in embeddings_dict if item["hardware_id"] == str(row["hardware_id"])),
                []
            )
        }
        for _, row in df.iterrows()
    ]
    for chunk in chunk_data(data, chunk_size):
        try:
            result = search_client.upload_documents(documents=chunk)
            print(f"Uploaded {len(chunk)} documents successfully.")
        except Exception as e:
            print(f"An error occurred during document upload: {e}")
            return None