DSXchange

Use StringToDecimal function in the transformer. It could be the case that the substring is being treated as character.
StringToDecimal(Table_Read.Source_Col[21,25])

DSFIELD1 : YEAR(TO_DATE(AUC_SALE_DATE,'yyyy-mm-dd'))
DSFIELD2 : WEEK(TO_DATE(AUC_SALE_DATE,'yyyy-mm-dd'))

Transformer Stage :

Column Derivation
DSFIELD1 : '-' : Str('0',2-len(DSFIELD2)):DSFIELD2

Goto Format tab-->Field Defaults-->choose property "Delimiter String"
Here you can enter more than one delimiter.

Refer to my solution :

viewtopic.php?t=152746

Read the input twice. Number the rows @INROWNUM.
Join COL1 to COL2 and COL2 to COL1 in lookup stage.
So there will be 4 cols col1,col2,col3,col4.
AUS,IND,IND,AUS,1,3
IND,AUS,AUS,IND,3,1
Then add Num1+Num2
AUS,IND,IND,AUS,4
IND,AUS,AUS,IND,4
Then do remove duplicate on Num.

For each row join the Key and Name from the main data with the value in the reference table. So instead of maintaining Name = 'P' maintain
Key V_SQL
1 'P'
Then in the lookup use Drop condition to drop the the incoming record if a join is unsuccessful.

read the sequential file with format :
Delimiter = comma
Quote = Double

This will separate the fields. Then the column with value "12,34" will contain 12,34. Then you can use Convert function to remove the ,
Convert(',','',DSLINK.COL)
This will yield 1234

It replaces multiple occurances of a sub-string in a string.

Please refer to my solution :

viewtopic.php?t=152746

First filter Extra from the data. Then in sequential transformer sort NTFCTN_ID ascending,SEQ_NUM ascending. Then use stage variables : svSeqNo initial value 0 : if input.svNTFCTN_ID = svNTFCTN_IDold then svSeqNo+1 else 1 svNTFCTN_IDold : input.svNTFCTN_ID svPivot : if mod(svSeqNo,2) = 0 then 1 else...

pivot them into rows

you could strip the delimited file in the before-job routine by invoking shell script.
sed "s/,//g" file.txt

Use a execute command stage in sequence job.
cat filename
Then use User Variables stage.
var : exec_command.output[1,len(exec_command.output) -1]
Pass this variable to your job as a parameter from Activity variables
user_variable.var

make the field with null value nullable Yes in the column metadata definition.

Read the timestamp field as varchar.
Then in transformer use conversion :
StringToTimestamp(COL,'%yyyy%mm%dd%hh%nn%ss.3')
This will convert the string to desired format of three decimal microseconds.