I have to find difference between alternate rows for a field like Row 2-1, Row 4-3. Not difference between all the consecutive rows. I used "diff" but it gives me difference between all the rows. Please suggest.

[BTCM]

TXName Period Value diffValue
tx1 Period 1 25

tx1 Period 2 14 -11
tx2 Period 1 12
tx2 Period 2 20 8

[javiergn]

Streamstats is your friend here.
This is what I would do:

| streamstats count as ID by TXName
| sort -limit=0 -ID
| eval ValueForDiff = if(ID=1, -Value, Value)
| eventstats sum(ValueForDiff) as DiffValue by TXName
| eval DiffValue = if(ID=1, null(), DiffValue)

Example (mycsv.csv contains the table you mentioned in your question):

| inputcsv mycsv.csv 
| streamstats count as ID by TXName
| sort -limit=0 -ID
| eval ValueForDiff = if(ID=1, -Value, Value)
| eventstats sum(ValueForDiff) as DiffValue by TXName
| eval DiffValue = if(ID=1, null(), DiffValue)
| table TXName, Period, Value, DiffValue
| sort -limit=0 TXName, Period

Output (see picture):