There is requirement that we have to reorder the features.

We have two type of cases one is general and other is specific.

In case of general, you can take the following steps
1- In each line check if fourth column has multiple feature and lcat="v" then
2- parse the features as key lcat_gend_num_pers_cm 
fs1=v_any_any_any_0 fs2=v_m_sg_any_yA
3- select second feature as first and first as 2nd.

<af='root,cat,gen,num,per,case,tam/cm,suff')>



input:
1 	calA 	<fs af='cala,v,any,any,any,,0,0'>|<fs af='cala,v,m,sg,any,,yA,yA'>

output:
1 	calA 	<fs af='cala,v,m,sg,any,,yA,yA'>|<fs af='cala,v,any,any,any,,0,0'>






examples:
General Case
1 	calA 	<fs af='cala,v,any,any,any,,0,0'>|#<fs af='cala,v,m,sg,any,,yA,yA'>


In case of specific, you can take the following steps

Specific Case
1	karake	#<fs af='kara,v,any,any,any,,kara,kara'>
		<fs af='karaka,v,any,sg,2,,e,e'>
		<fs af='karaka,v,any,sg,3,,e,e'>
		<fs af='karaka,v,m,pl,any,,yA,yA'>



2 	आने 	#<fs af='आ,v,m,pl,any,,ना,nA'>
		<fs af='आन,v,any,sg,2,,ए,e'>
		<fs af='आन,v,any,sg,3,,ए,e'>
		<fs af='आन,v,m,pl,any,,या,yA'>

# feature should be first