Empty pandas dataframe when querying from redshift using psycopg2

Question

I would like to read data from redshift table and load it to dataframe and perform transformations. I used psycopg2 to connect to redshift and used pandas read_sql to query the table as below

con = psycopg2.connect("dbname=sales host=redshifttest-xyz.cooqucvshoum.us-west-2.redshift.amazonaws.com port=5439 user=master password=secret")
cur = con.cursor()
sql = "select * from dtw.rpt_account_transfer_hist where transfer_date>=2020-07-01;"
df  = pd.read_sql(sql, con)

I see an Empty Dataframe but data exist when I query the database. When I print the schema everything is non-null object .

I parameterized the transfer_date as below and tried again. This time the whole data set is returned without any filter being applied. Not sure where I'm missing. I tried cast in the sql query itself but it returned an empty dataframe. Any leads please.

curr_dt = datetime.strftime(datetime.now() - timedelta(3), '%Y-%m-%d')
sql = "select * from dtw.rpt_account_transfer_hist where transfer_date>=" +str(curr_dt)+";"
df  = pd.read_sql(sql, con)

The data in redshift table is like below with datatype as varchar for col1 , col2 , col4 , col5 and date for transfer_date .

col1    col2   transfer_date col4    col5
6052148 670018  2020-07-13  640033  6052148
5260969 640737  2020-07-11  640033  5260969
4778065 610050  2020-07-11  610017  4778065
7942224 690020  2020-07-11  690032  7942224
5260969 640737  2020-07-10  640033  5260969
4778065 610050  2020-07-10  610017  4778065
7942224 690020  2020-07-10  690032  7942224
5073022 640601  2020-07-09  640679  5073022
0309991 640601  2020-07-09  640729  0309991

Answer 1

I think you're missing single quotes around the date, try with this:

sql = "select * from dtw.rpt_account_transfer_hist where transfer_date>='2020-07-01';"

Answer 2

Sounds a bit weird, I haven't changed anything and it started working.