我从网页上读取了很多json。由于出现内存错误,无法一次读取。所以我正在尝试使用ijson库流式传输它。问题是我丢失了json的结构,所以我无法正确,有序地获取数据。

我能够读取键值对等,但是没有结构了。

这是我的代码:

parser = ijson.parse(urllib.request.urlopen('https://data.medicaid.gov/resource/4qik-skk9.json?$limit=5'))
for prefix, event, value  in parser:
    print(str(prefix) +" "+str(event)+ " "+str(value))


这是json的结构:

[
    {
        "package_size_code": "60",
        "fda_ther_equiv_code": "NR",
        "fda_application_number": "204153",
        "clotting_factor_indicator": "N",
        "year": "2018",
        "fda_product_name": "LUZU Cream 1% 60gm",
        "labeler_name": "MEDICIS DERMATOLOGICS, INC.",
        "ndc": "99207085060",
        "product_code": "0850",
        "unit_type": "GM",
        "fda_approval_date": "2013-11-14T00:00:00",
        "market_date": "2014-03-14T00:00:00",
        "pediatric_indicator": "N",
        "package_size_intro_date": "2014-03-14T00:00:00",
        "units_per_pkg_size": "60000",
        "labeler_code": "99207",
        "desi_indicator": "1",
        "drug_category": "S",
        "quarter": "3",
        "cod_status": "3"
    },
    {
        "package_size_code": "60",
        "fda_ther_equiv_code": "AB",
        "fda_application_number": "21758",
        "clotting_factor_indicator": "N",
        "year": "2018",
        "fda_product_name": "VANOS CREAM .1%",
        "labeler_name": "MEDICIS DERMATOLOGICS, INC.",
        "ndc": "99207052560",
        "product_code": "0525",
        "unit_type": "GM",
        "fda_approval_date": "2005-02-11T00:00:00",
        "market_date": "2005-02-21T00:00:00",
        "pediatric_indicator": "N",
        "package_size_intro_date": "2005-02-21T00:00:00",
        "units_per_pkg_size": "60000",
        "labeler_code": "99207",
        "desi_indicator": "1",
        "drug_category": "I",
        "quarter": "3",
        "cod_status": "3"
    },
.
.
.
.
]


我从使用ijson得到的输出是:

 start_array None
item start_map None
item map_key clotting_factor_indicator
item.clotting_factor_indicator string N
item map_key cod_status
item.cod_status string 4
item map_key desi_indicator
item.desi_indicator string 1
item map_key drug_category
item.drug_category string I
item map_key fda_application_number
item.fda_application_number string 50007
item map_key fda_approval_date
item.fda_approval_date string 1990-09-30T00:00:00.000
item map_key fda_product_name
item.fda_product_name string DOXYCYCLINE HYCLATE 100MG CAP
item map_key fda_ther_equiv_code
item.fda_ther_equiv_code string AB
item map_key labeler_code
item.labeler_code string 59762
item map_key labeler_name
item.labeler_name string PFIZER, INC.
item map_key market_date
item.market_date string 1990-09-30T00:00:00.000
item map_key ndc
item.ndc string 59762369001
item map_key package_size_code
item.package_size_code string 01
item map_key package_size_intro_date
item.package_size_intro_date string 2015-05-01T00:00:00.000
item map_key pediatric_indicator
item.pediatric_indicator string N
item map_key product_code
item.product_code string 3690
.
.
.


我想以某种方式知道哪些数据属于哪个对象。有点像索引,但是我不确定该怎么做。

最佳答案

也许,更好的主意是使用ijson.items

item_gen = ijson.items(urllib.urlopen('https://data.medicaid.gov/resource/4qik-skk9.json?$limit=5'), 'item')

那么您可以在for循环或任何itertools中遍历item_gen。例如:

for item in item_gen:
    print(item)


我使用了Python 2,但我想它在Python 3中非常相似。

我从here那里得到了这个主意。

关于python - 如何解析json流中的对象,我们在Stack Overflow上找到一个类似的问题:https://stackoverflow.com/questions/53888358/

10-11 08:42