Solusi saya adalah membuka file BSON secara manual (dengan python), temukan dokumen besar dan hapus beberapa bagiannya, lalu tulis objek BSON ke file BSON baru dan muat file BSON yang diedit, yang berhasil disimpan ke mongo.
Ini tidak memenuhi keinginan saya untuk dapat memuat db yang dibuang ke sistem tanpa mengubahnya!
Python3:
import bson
from pprint import pprint
def get_bson_data(filename):
with open(filename, "rb") as f:
data = bson.decode_all(f.read())
return data
def report_problematics_documents(data):
problematics = []
for item in data:
if is_too_big(item):
print(item)input("give me some more...")
input("give me some more...")
problematics.append(item)
print(f"data len: {len(data)}")
print(f"problematics: {problematics}")
print(f"problematics len: {len(problematics)}")
def shrink_data(data):
for i, item in enumerate(data):
if is_too_big(item):
data[i] = shrink_item(item) # or delete it...
print(f"item shrinked: {i}")
def write_bson_file(data, filename):
new_filename = filename
with open(new_filename, "wb") as f:
for event in data:
bson_data = bson.BSON.encode(event)
f.write(bson_data)
def is_too_big(item):
# you need to implement this one...
pass
def shrink_item(item):
# you need to implement this one...
pass
def main():
bson_file_name = "/path/to/file.bson"
data = get_bson_data(bson_file_name)
report_problematics_documents(data)
shrink_data(data)
report_problematics_documents(data)
new_filename = bson_file_name + ".new"
write_bson_file(data, new_filename)
print("Load new data")
data = get_bson_data(new_filename)
report_problematics_documents(data)
if __name__ == '__main__':
main()
