MVSFORUMS.com

[atakella]

Hi all,
I have this requirement of loading 20 byte data in the form of
(-123.456988231876000) from PS file into DB2 column which is defined as FLOAT datatype.(Double Precision occupies 8 bytes)

I tried various possibilities listed as below:
1. Loaded the data as is W 20 A. Failed with S0C7.
2. Conver the data from W 20 A to another variable of W 18 N/W 18 N 15, and then load that data to DB2. But easytrieve thrown error "B204 SQL-INVALID HOSTVAR DATATYPE 836"
3. Tried to move the data W 20 A to packed variable W 9 P 14/ W 10 P 16 and then load that data to DB2. Getting error "B055 INVALID LENGTH, TYPE OR DECIMAL PLACES.
4. Tried to move the data from W 20 A to numeric variable W 18 N 14; which in turn has been moved to another variable of W 10 P 14. Failing with S0C7.
5. Tried to use DB2 CAST scalar function with W 20 A variable. Fetch worked fine with (CAST AS CHAR(20)) but while updating the (CAST AS DOUBLE) failing with -301 SQLCODE.

Is there any way to accomplish this task? Am I missing something in this entire process??
===============================================

Then, I edited my input file to remove decimal point '.' Data in the form of
(-123456988231876000) from PS file into DB2 column which is defined as FLOAT datatype. Now data only 19 bytes.
I had defined MASK while loading the data to numeric variable as
'ZZZ.99999999999999-'. Still unable to load the data as is. Getting S0C7.
===============================================

Then, I edited my input file to remove -ve sign. This time I was able to load the data into table. DB2 assumed my decimal point correctly but it couldn't recognize the -ve sign. So my entire file was loaded to DB as +VE floating point numbers which I don't want.

So to summarize is it that only +ve Float numbers can be handled through Easytrive to load to DB2?
Is there anyway we can load Float data to DB2 at all?

Please let me know if you need any additional info. I have been trying numerous possibilities but couldn't crack it. so thought of taking big brains help!!

-Akella

[kolusu]

atakella,

Define the Easytrieve variable as

W 22 A

and move the value as -123.45678901234567890

and in the load card assuming that float data starts at position 35 , use the following control card. External function makes DB2 to take it in numeric format and convert it into float format

[atakella]

I am sorry if I carried the impression of loading data to DB2 through easy trieve using LOAD CARD. It's basically updation of couple of fields on existing DB2 record. And those two fields are Float data types. And I am facing this problem while updating the record. I get the data through PS, after initial validations and compares, I will try to update these fields to DB2.
So guess, I might not be able to use LOAD CARD as such. I am not quite well into easytrieve, correct me if I am mistaken.

Thanks and awaiting your response.

[kolusu]

atakella,

try this (untested)

[atakella]

Tried....

[kolusu]

atakella,

You cannot move a alpha-numeric variable directly into a packed decimal field. You need to redefine it and move it.

Take a look at the example here

http://mvsforums.com/helpboards/viewtopic.php?p=24520#24520

You have the data in similar format. First move that value to a zoned decimal format removing the decimal and sign as shown in the example and then move that value to packed decimal field.

Kolusu

[atakella]

Thanks a lot. With the link given, I was able to transform string expression of float value to numeric and then to Packed Decimal and finally to DB2 table as FLOAT(Double Precision).

However, one thing I observed was...

[kolusu]

atakella,

Hmm I think you made a mistake of defining the fields as it increased the decimals to 16 instead of 15. How did you define your Numeric and Packed decimal fields? You need to define the numeric variable as 20 bytes and the packed decimal as 10 bytes.

Kolusu
_________________
Kolusu
www.linkedin.com/in/kolusu

[atakella]

Please find the declaration I have used.

[atakella]

Just to add above, I couldn't define numeric as 20 bytes. The max I could define for numeric is 18 bytes

System throws error for

[kolusu]

atakella,

Did you get the right values for negative float values? You defined the sign field as 1 byte field and you are checking for 2 bytes? why are you even checking for -1 ? when you just need to check if it is just a minus symbol?

Why do you have a MASK on a working storage field? Unless you are writing a report that mask does nothing.

You cannot define a 20 byte numeric field however you only have 18 byte of data after the removal of decimal point and the sign.

Also learn to initialize the working storage fields to zero before you manipulate them.

So try this piece of code(untested)

[atakella]

Apologies for late reply. I was on vacation.

1. It was my typo error. I did check for minus symbol"-" only in the actual piece of code

2. I masked the working storage field because I am printing the values to SYSOUT for testing purpose. Will remove once I am through successful load/update of float data

3. Agree. 18 byte of data after removal of decimal point and the sign.

4. I did initialize the working storage fields to ZERO, just that I din't paste it over here bcoz I wanted only the declaration and manipulation with them.

5. Tried your piece of code, and

[kolusu]

[atakella]

But that is what I tested initially(see my code snippet posted on (14th June).

And I was getting 16 digits after decimal point. And 16th digit is either 1 or 2 numbers added or substracted.