在SQLAlchemy数据库表中创建并追加到列表

时间:2018-09-18 16:35:21

标签: python sql sqlalchemy

我正在学习SQLAlchemy并被卡住。 我有一个SQL表(table1)有两个字段:'name'和'other_names'

我有一个包含两列的excel文件:

first_name alias   
paul   patrick
john   joe
simon  simone
john   joey
john   jo

我想将excel文件读到我的table1中,使其看起来像这样(即同一行的所有别名都在一行上):

paul    patrick
john    joe,joey,jo
simon   simone

这是我试图做的想法。我尝试过的代码(带有注释):

for line in open('file.txt', 'r'): #for each line in the excel file
        line = line.strip().split('\t') #split each line with a name and alias
        first_name = line[0] #first name is the name before the tab
        alias = line[1] #alias is the name after the tab
        instance = 
        Session.query(session,tbs['table1'].name).filter_by(name=first_name) #look through the database table, by name field, and see if the first name is there 
        list_instance = [x[0] for x in instance] #make a list of first names already in database table
        if first_name not in list_instance: #if the excel first name is not in the database table
              alias_list = [] #make an empty list
              alias_list.append(alias) #append the alias
              name_obj = lib.get_or_create( #small function to make db object
              session,
              tbs["table1"],
              name = first_name, #add first name to the name field
              other_names = alias_list # add alias list to the other_names field
            )


       elif first_name in list_instance: #elif first name already in db
             alias_list.append(alias) #append the alias to the alias list made above
             name_obj = lib.get_or_create(
             session,
             tbs["table1"],
             name = first_name,
             other_names = alias_list #create object as before, but use updated alias list
    )

问题是我可以使上面的代码无错误运行,而且输出不是附加列表,它只是一个数据库表,看起来像excel文件;即

name   alias
paul   patrick
john   joe
simon  simone
john   joey
john   jo

有人可以指出我要去哪里,特别是如何修改此代码?如果问题不清楚,请让我知道,我试图将其作为一个简单的例子。具体来说,我如何初始化并添加到列表中作为SQLalchemy db表中的字段条目。

更新1:我已根据以下同类建议更新了我的代码。但是我仍然有问题。这是完整的目标,代码和测试文件: 目的:

我在数据库中有一个表(进入表的测试文件请参见下文),该表有两个字段,名称(拉丁名称,例如智人)和其他名称(常见名称,例如人,人)。我想更新表中的字段(其他名称),所以不要:

Rana rugosa human   
Rana rugosa man 
Rana rugosa frog    
Rana rugosa cow

我有:

Rana rugosa human,man,frog,cow

test_data文件如下所示:

origin_organism        common_name         tested_organism
Rana rugosa            human                -
Rana rugosa            man                  -
Rana rugosa            frog                 homo sapiens
Rana rugosa            cow                  Rana rugosa
Rana rugosa            frog                 Rana rugosa
Rana rugosa            frog                 -
Rana rugosa            frog                 -
Rana rugosa            frog                homo sapiens
-                      -                   -
-                      -                   homo sapiens
-                      -                   -
-                      -                   -
-                      -                   -
-                      -                   -
streptococcus pneumoniae    -              -

代码:

import sys 
from sqlalchemy.orm  import * 
from sqlalchemy  import * 
from dbn.sqlalchemy_module  import lib 
import pd

engine = lib.get_engine(user="user", psw="pwd", db="db", db_host="111.111.111.11")
Base = lib.get_automapped_base(engine)
session = Session(engine)
tbs = lib.get_mapped_classes(Base)
session.rollback()
df = pd.read_excel('test_data.xlsx', sheet_name = 'test2')




for index, row in df.iterrows():  
    origin_latin_name = row['origin_organism'].strip().lower()
    other_names_name = row['common_name'].strip().lower()
    tested_species = row['tested_organism'].strip().lower()


if origin_latin_name not in [None, "None", "", "-"]:
    instance = [x[0] for x in Session.query(session,tbs['species'].name).filter_by(name=origin_latin_name).all()]
    if origin_latin_name not in instance:
        origin_species = lib.get_or_create(
            session,
            tbs["species"],
            name = origin_latin_name,
            other_names = other_names_name
        )

    elif origin_latin_name in instance:
        other_names_query = Session.query(session,tbs['species'].other_names).filter_by(name=origin_latin_name)
        other_names_query_list = [x for x in other_names_query]
        original_list2 = list(set([y for y in x[0].split(',') for x in other_names_query_list]))
        if other_names_name not in original_list2:
            original_list2.append(other_names_name)
            new_list = ','.join(original_list2)
            new_names = {'other_names':','.join(original_list2)}

        origin_species = lib.get_or_create(
            session,
            tbs["species"],
            name = origin_latin_name,
            other_names = new_list
        )

elif语句中的部分无效。我遇到了两个问题:

(1)我得到的最新错误: NameError:名称“ new_list”未定义

(2)我遇到的另一个错误是我还有另外一张桌子

map1 = lib.get_or_create(
    session,
    tbs["map1"],
    age_id_id = age,
    name_id_id = origin_species.id
    )

...并且它说无法找到origin_species,但是我认为这与elif语句相关联,以某种方式无法正确更新origin_species对象。

如果有人可以帮助,我将不胜感激。

2 个答案:

答案 0 :(得分:3)

简单的错误。您没有列出清单。我不确定为什么它们会以不同的行结尾,但是,我将更改以下内容,因为目前我看不到将名称拆分到列表中的位置,我所看到的只是使用以下命令将字符串分配给列表追加。

alias_list = alias.split(',')

也可能是:

alias_list = line[1].split(',')

输出:

alias_list:    ['Name1','Name2','Name3']

当前您的代码输出:

alias_list = ['Name1,Name2,Name3']

虽然从技术上讲它是按数据类型的列表,但对于您要使用的方式而言,它是毫无价值的列表。这是因为alias_list[0]将返回整个字符串,而不是'Name1'

警告语:

您的代码不必要地创建了一个列表。您的数据库中不需要列表,使用读取excel文件时评估的字符串就可以轻松实现自己的目标。

IMHO应该做的是将名称字符串存储为一个完整的字符串,然后,如果需要查询某人的别名,那么可以在另一侧拆分该字符串吗?

答案 1 :(得分:0)

如果对熊猫数据框使用groupby语句,则可以轻松地做到这一点。代码未经测试,让我知道是否存在错误

import sys 
from sqlalchemy.orm  import * 
from sqlalchemy  import * 
from dbn.sqlalchemy_module  import lib 
import pd

engine = lib.get_engine(user="user", psw="pwd", db="db", db_host="111.111.111.11")

###Get test data
added_df = pd.read_excel('test_data.xlsx', sheet_name = 'test2')

###Get current database as a pandas object
my_current_df = pd.read_sql_table(
        "table1", engine)

#Now group by origin organism
gb = added_df.groupby('origin_organism')

#For every member that had the same origin organism, return it as a comma seperated list into a new dataframe
df = pd.DataFrame(gb.alias.apply(lambda x: ",".join(x)))

#Now append that dataframe onto the one that came from the database
my_current_df.append(df)

#NOw return that dataframe back to the database
my_current_df.to_sql(
        name='table1',
        con=engine,
        if_exists='replace')