The idea of how to delete lines from CSV file itself is more or less easy to understand - but you used your CSV file in such a different way than the most appropriate way of using it, which - at least I - I went through the question a couple of times without answering just thinking about the hassle of trying to explain the most appropriate way.
But here we go:
First - your idea is right: there is no way to "modify an existing file": it is always necessary to create a new file and save all the content desired to the file. Of course in general it is best to do this in 3 steps: (1) open the old file; (2) write the new data in a file with a different name; (3) finally remove the original file, and rename the new file to take its place. These 3 simple steps prevent that if your program stops in the middle by an error of any nature, your data is lost - at any time you have either the original file, or the new file already completed.
From there on the peculiar things of this file start: the normalpara Mum CSV file is that it is a single table, with all the lines having the same structure. In this case, you have created two types of different lines, meaning different things - the simplest, to keep the data in CSV in this case, is to keep all the data in each line (ie: denormalized).
But also thinking: will putting more features on top of a CSV structure solve your problem? If you want to associate "critical" type objects with "movie" type objects, you have a relationship - suddenly a relational database is best for you there.
If this file is never going to get big compared to the memory of the computer you are going to run (think that a modern PC has about 2GB of memory, and the entire Bible in text takes about 3MB - that is, if you have the equivalent of the Bible (~1000 pages in small print) of textual data, will be occupying about 0.2% of the computer’s memory: which means that most likely you may have a program that works all the time with all the data in memory, and only "save" the content when it is convenient. In this case, you have to save in a format that can be read back in another program execution: it can be either an unconventional CSV, or what you have, or it can be a serialized file like "JSON": a syntax similar to the dictionaries we use in Python, which can be read and written directly to a text editor, but which has two advantages in this case: all data can be read from the file, or written to it, in a single function call, and, more importantly: in a JSON file you can preserve the hierarchical structure of your information - and take advantage to use the same structure in memory.
Let’s work with this representation of your data in a list of dictionaries in memory: this will allow you to have a program that does more operations with your objects - and, to keep me on the subject of the question, I will put functions to read your current CSV file to this structure, and save this structure to a format like the one you have:
import csv
import json
import sys
def le_dados_csv(nome_do_arquivo):
dados = []
with open(nome_do_arquivo) as arq:
leitor = csv.reader(arq, delimiter=";")
critico = None
for linha in leitor:
if linha[0] == "C": # Dados de um novo crítico
critico = {}
critico["codigo"] = linha[1]
critico["nome"] = linha[2]
critico["email"] = linha[3]
critico["filmes"] = []
dados.append(critico)
else: # A informação nesta linha é sobre um filme do último crítico
filme = {}
filme["titulo"] = linha[1]
filme["nota"] = float(linha[2])
# acrescenta as ifnormações sobre este filme a
# lista de filmes do último crítico
critico["filmes"].append(filme)
return dados
def grava_dados_csv(nome_do_arquivo, dados):
with open(nome_do_arquivo, "wt") as arq:
escritor = csv.writer(arq, delimiter=";")
for critico in dados:
escritor.writerow(("C", critico["codigo"], critico["nome"], critico["email"]))
for filme in critico["filmes"]:
escritor.writerow(("F", filme["titulo"], "{:.01f}".format(filme["nota"])))
def grava_dados_json(nome_do_arquivo, dados):
"""Grava os dados do programa como uma estrutura json"""
with open(nome_do_arquivo, "wt") as arq:
json.dump(dados, arq, indent=4, separators=(',', ': '))
def le_dados_json(nome_do_arquivo):
"""Grava os dados do programa como uma estrutura json"""
with open(nome_do_arquivo) as arq:
dados = json.load(arq)
return dados
def remove_critico(dados, nome_critico):
for indice, critico in enumerate(dados):
if critico["nome"] == nome_critico:
del dados[indice]
break
else:
# A clausula else de um "for" em Python é executada se o "for"
# terminou sem ser por causa de um comando "break".
print("Critico {} não encontrado".format(nome_critico), file=sys.stderr)
# E por fim, uma função "principal" que faz o que você descreve
# na pergunta - acreido que fica fácil você ampliar seu programa
# a partir daqui:
def principal():
dados = le_dados_csv("arquivo.csv")
remove_critico("Cleiton")
grava_dados_csv("arquivo.csv")
# Executar a função principal apenas se este arquivo Python
# for executado como programa principal.
# Isso permite que outros rquivos .py possam importar
# este arquivo e usar as funções de leitura e escrita
# normalmente
if __name__ == "__main__":
principal()
Answered here.
– Sidon