How to query data by time period in PB

Hello, date(string(em,'yyyymmdd') changed to. date(string(em em 1 requires a date like 2015-01-05.

datet dt_test

dt_test = '2015-01-02'

string(dt_test ,'yyyymmdd'That's how it's written.

Data window name dw a

The data window column name is a date

The filter statement should be.

If you take the time type:

datetime ldt_date

ldt date=dw line number].

If you take the character type:

string ls_date

ls_date="2013/12/24"or ls date=string(date(ldt date),'yyyy/mm/dd')

dw_"a_date>'"+ls_date+"'")dw_

Have you built a dw??

The easiest way is to create conditional parameters in dw, direct dw If you don't have conditional parameters, you have to write **. Write below in the button string ls oldsql, ls newsqlls oldsql = dw

ls_newsql = ls_oldsql + "where column name ='" + sle_ +"'"

If you have a condition in DW, use ls newsql = ls oldsql +"and Column Name='" + sle_ +"'"

dw_dw_dw_

1.Add search conditions to the data window object. Let's say your column is a glyph and you can write dw in the button

2.button to write the following. dw_

dw dw column name ="'"+trim(sle_"'")dw_

3.Dynamic SQL implementation.

It is necessary to create a clustered index on the ID first.

If the id is sequential, it is quick for the upstairs to use id to qualify the range of statements.

If the id is not continuous, the landlord's statement uses not in, which is generally very inefficient.

You can change the statement to:

select top m *

from table

where id> (select max(id) from (select top n id from table order by id) a)

order by id

This can improve some performance to a certain extent.

For example, I have more than 3 million records in my table, and it takes more than 9 seconds to execute the landlord's statement, and more than 3 seconds to use my statement.

This is not difficult, two knowledge points.

1 is the content acquisition of the text box SLE.

2 is a conditional update for DW.

First of all, you can do it in the OK button, name example ls name = sle.

Of course, you can also make a judgment, if ls name =''or isnull(ls name )then indicates whether to obtain this information.

The second is the DW update, which can pass the conditions just obtained to the DW as parameters, and set the transfer parameters when designing the DW.

You can also create a DW with a dynamic SQL statement, and write the parameters just now as a specific method for the where condition, which are explained in detail in the book, and you can just find the knowledge points according to the idea.