Python词典：将多个值列表合并为一个唯一值列表

Coll_id：条目可以是单个收集者,也可以是组

Participant_Coll_id：如果Coll_id是单个收集器,则该值将为null.如果Coll_id是一个组,那么该组中的每个参与者将只有一行.

样本在这里：

Coll_id,Participant_Coll_id<br>
ARA,ARG  
ARA,RAT  
ARG,NULL  
BRSAR,SGMB  
BRSAR,SANTM  
BRSAR,CRSR  
BRSAR,RAT  
CRSR,NULL  
DBY,NULL  
HZIE,NULL  
RAT,NULL  
SANTM,NULL  
SGMB,NULL  
ARG,NULL  
DRS,CRSR  
DRS,RAT  
DRS,ARG

对于每个收集器(coll_id),我正在尝试创建他们收集的所有其他收集器的列表.我试图将代码拉到一起来执行以下操作,它现在非常接近：

#This is giving me a dictionary with each COLL_ID having a list of PARTICIPANT_COLL_IDs

with open('colls_mv1.csv','r') as f:
    reader = csv.DictReader(f,['COLL_ID','PARTICIPANT_COLL_ID'])
    data1 = defaultdict(list)

    for line in reader:
        data1[line['COLL_ID']].append(line['PARTICIPANT_COLL_ID'])


#And this is giving me a dictionary with each PARTICIPANT_COLL_ID having a list of COLL_IDs
with open('colls_mv1.csv','PARTICIPANT_COLL_ID'])
    data2 = defaultdict(list)

    for line in reader:
        if line['PARTICIPANT_COLL_ID'] != 'NULL':
            data2[line['PARTICIPANT_COLL_ID']].append(line['COLL_ID'])

dict3 = {k: [data1[i] for i in v] for k,v in data2.items()}

print dict3

我得到以下输出：

{'SGMB': [['SGMB','SANTM','CRSR','RAT']],'CRSR': [['SGMB','RAT'],['CRSR','RAT','ARG']],'RAT': [['ARG',['SGMB','PARTICIPANT_COLL_ID': [['PARTICIPANT_COLL_ID']],'ARG': [['ARG','SANTM': [['SGMB','RAT']]}

我想要的是将值列表合并为每个键,删除重复项并从值列表中删除键：

{'SGMB': ['SANTM','CRSR': ['SGMB','ARG'],'RAT': ['ARG','SGMB','CRSR'],'ARG': ['RAT','SANTM': ['SGMB','RAT']}