使用 AWS athena 的时间戳差异

Question

I am running an SQL query using AWS Athena in my jupyter notebook as follows. It involves computing differences between timestamps as follows.

query_demog = """


select ad.subject_id, ad.hadm_id, i.icustay_id ,
       date_diff('second', timestamp '1970-01-01 00:00:00', ad.admittime) as admittime,
       date_diff('second', timestamp '1970-01-01 00:00:00', ad.dischtime) as dischtime,
       ROW_NUMBER() over (partition by ad.subject_id order by i.intime asc) as adm_order,
       case when i.first_careunit='NICU' then 5
            when i.first_careunit='SICU' then 2
            when i.first_careunit='CSRU' then 4
            when i.first_careunit='CCU' then 6
            when i.first_careunit='MICU' then 1
            when i.first_careunit='TSICU' then 3
       end as unit,
       date_diff('second', timestamp '1970-01-01 00:00:00', i.intime) as intime,
       date_diff('second', timestamp '1970-01-01 00:00:00', i.outtime) as outtime,
       i.los,
from mimiciii.admissions ad,
     mimiciii.icustays i,
     mimiciii.patients p
where ad.hadm_id=i.hadm_id and p.subject_id=i.subject_id 
order by subject_id asc, intime asc

"""

It works fine. Now when I include another line having similar timestamps differences, I get an error.

query_demog = """


select ad.subject_id, ad.hadm_id, i.icustay_id ,date_diff('second', timestamp '1970-01-01 00:00:00', ad.admittime) as admittime, date_diff('second', timestamp '1970-01-01 00:00:00', ad.dischtime) as dischtime, ROW_NUMBER() over (partition by ad.subject_id order by i.intime asc) as adm_order, case when i.first_careunit='NICU' then 5 when i.first_careunit='SICU' then 2 when i.first_careunit='CSRU' then 4 when i.first_careunit='CCU' then 6 when i.first_careunit='MICU' then 1 when i.first_careunit='TSICU' then 3 end as unit,  date_diff('second', timestamp '1970-01-01 00:00:00', i.intime) as intime, date_diff('second', timestamp '1970-01-01 00:00:00', i.outtime) as outtime, i.los,

 EXTRACT(EPOCH FROM (i.intime-p.dob)::INTERVAL)/86400 as age





from mimiciii.admissions ad, mimiciii.icustays i, mimiciii.patients p

where ad.hadm_id=i.hadm_id and p.subject_id=i.subject_id 

order by subject_id asc, intime asc

"""

The inclusion of the line EXTRACT(EPOCH FROM (i.intime-p.dob)::INTERVAL)/86400 as age creates an error as follows.

An error occurred (InvalidRequestException) when calling the StartQueryExecution operation: line 3:37: mismatched input ':' expecting {'.', ')', '[', 'AT', '+', '-', '*', '/', '%', '||'} unable to rollback

I don't know why the inclusion of EXTRACT(EPOCH FROM (i.intime-p.dob)::INTERVAL)/86400 as age creates an error

Answer 1

to_unixtime()内置应该工作：

to_unixtime(i.intime-p.dob)/86400 as age