Combining 2 Tables, Want to keep all Non 0 Bit on DUPLICATE KEY

Question

I have a table with 1000 columns (yes its normalized), that is storing Biiig Daaata!

I need to Insert Update new data as it becomes available overnight, and figured the fastest way to do that is to take the data, generate a CSV file and insert into table 2.

Table 1              Table2
(Real Data)         (Data to be merged)

Then do something like:

INSERT INTO blueData b 
SELECT * 
FROM rawData r 
ON DUPLICATE KEY 
  UPDATE _____

This is where I draw a blank, basically all the fields are bit fields (except PK) that is 1 or 0. I want to keep all the fields with 1 in both tables, the rest can stay zero.

Re-Explained, the Table2 will have new data that relates to table 1. I need to set the new bit fields from table 2 to the corresponding PK on table 1 without removing its 1 bit fields. But because it generates a CSV the 0's are being generated in the columns that arnt new for the data.

Table 1:

Row1:
PK: abc
BI: 1
BG: 0 
BD: 0
BF: 0

Table 2:

Row1:
PK: abc
BI: 0
BG: 1
BD: 1
BF: 0

Needs to end up like: Table1

Row1:
PK: abc
BI: 1
BG: 1
BD: 1
BF: 0

Hopefully this all makes sense :D

I need help writing the update part on duplicate key, only deal is with 1000 columns be nice if I didn't have to write a giant SQL statement that does if's on every column (will generate one in Python if I have to) but would make query huge!

Answer 1

Try use logical OR

INSERT INTO Table1
SELECT * FROM Table2 b 
ON DUPLICATE KEY UPDATE
 Table1.BI = (Table1.BI OR b.BI),
 Table1.BG = (Table1.BG OR b.BG),
 Table1.BD = (Table1.BD OR b.BD),
 Table1.BF = (Table1.BF OR b.BF)

It will give you this in Table1

|  PK | BI | BG | BD | BF |
---------------------------
| abc |  1 |  1 |  1 |  0 |

SQLFiddle