在Julia循环中处理JSON字段双值或空值的问题

问题根源 #

你的代码存在两个核心问题：

1. 误用了R的NA，Julia中处理缺失值的标准是missing；
2. 当classeCooperativa为nothing时，直接嵌套调用get会尝试从非字典类型取值，导致报错。

修正方案 #

以下是修改后的完整代码，关键部分做了注释说明：

using HTTP, JSON, DataFrames

# 提前初始化空DataFrame，明确列类型更规范
info_gerais_coop = DataFrame(
    cnpj_coop = Int[],
    nome_coop = String[],
    naturezaJuridica = Union{String, Missing}[],
    classe = Union{String, Missing}[],
    situacao = Union{String, Missing}[],
    regimeEspecial = Union{Any, Missing}[],
    logradouro = Union{String, Missing}[],
    complemento = Union{String, Missing}[],
    bairro = Union{String, Missing}[],
    municipio = Union{String, Missing}[],
    uf = Union{String, Missing}[],
    cep = Union{String, Missing}[],
    endereco_eletronico = Union{String, Missing}[],
    email = Union{String, Missing}[],
    segmento_prudencial = Union{Any, Missing}[]
)

cnpj_list = [3795072, 1439107]

for cnpj in cnpj_list
    url = "https://www3.bcb.gov.br/informes/rest/pessoasJuridicas?cnpj=$cnpj"
    response = HTTP.get(url)
    dcoop = JSON.parse(String(response.body))

    # 先提取地址对象，避免重复调用get
    endereco = get(dcoop, "endereco", missing)
    
    # 安全提取classe字段：先判断classeCooperativa是否有效，再取值
    classe = let cc = get(dcoop, "classeCooperativa", missing)
        ismissing(cc) || cc === nothing ? missing : String(get(cc, "nome", missing))
    end

    # 构建单行数据
    row = (
        cnpj_coop = cnpj,
        nome_coop = dcoop["nome"],
        naturezaJuridica = String(get(dcoop, "naturezaJuridica", missing)),
        classe = classe,
        situacao = String(get(dcoop, "situacao", missing)),
        regimeEspecial = get(dcoop, "regimeEspecial", missing),
        logradouro = ismissing(endereco) ? missing : String(get(endereco, "logradouro", missing)),
        complemento = ismissing(endereco) ? missing : String(get(endereco, "complemento", missing)),
        bairro = ismissing(endereco) ? missing : String(get(endereco, "bairro", missing)),
        municipio = ismissing(endereco) ? missing : String(get(get(endereco, "municipio", missing), "nome", missing)),
        uf = ismissing(endereco) ? missing : String(get(get(endereco, "municipio", missing), "siglaEstado", missing)),
        cep = ismissing(endereco) ? missing : String(get(endereco, "cep", missing)),
        endereco_eletronico = ismissing(endereco) ? missing : String(get(endereco, "email", missing)),
        email = ismissing(endereco) ? missing : String(get(endereco, "email", missing)),
        segmento_prudencial = get(dcoop, "segmentoPrudencial", missing)
    )

    # 将行添加到DataFrame，比循环vcat高效
    push!(info_gerais_coop, row)
end

额外优化建议 #

• 推荐替换JSON包为JSON3：JSON3解析速度更快，返回带类型的结构体，能更安全地访问字段，减少类型不确定的问题；
• 提前定义DataFrame列类型：避免自动推断类型带来的错误，处理大量数据时更稳定；
• 直接遍历cnpj_list而非索引：代码更简洁易读，符合Julia的惯用写法。

内容的提问来源于stack exchange，提问作者RxT