Collapse rows to lowest complete row in each set - r

Question

I am cleaning a gigantic dataset, whcih came from using tabulizer() on a PDF.
The columns are correctly delineated, but I have many rows where one cell in the original was gigantic, and tabulizer read it as several rows, with all cells blank except the large one. I need to collapse the data frame so that rows are collapsed "down" to the lowest complete row.
Here is a sample of what the data look like:
As you can see, the column that these "extra rows" appear in varies by row (in one case it's species, in other cases it's area.of.operation. I want to collapse them to complete rows, such that row 1 remains intact, row 2 is actually rows 2:6 collapsed, row 7 is intact, etc. I don't even know if R is the best tool to use for this, but I would love if there is a dplyr solution. Example dataframe below.
Thank you in advance.
mydata <- structure(list(X = c(1L, 2L, 3L, 4L, 5L, 6L, 7L, 11L, 12L, 13L,
17L), target.species = structure(c(4L, 1L, 1L, 1L, 1L, 5L, 4L,
1L, 1L, 2L, 3L), .Label = c("", "hake", "hake, southern", "rosefish",
"squid, cuttlefish,"), class = "factor"), gear = structure(c(2L,
1L, 1L, 1L, 1L, 1L, 2L, 1L, 1L, 3L, 2L), .Label = c("", "trawl, bottom",
"trawl, midwater"), class = "factor"), number.boats = structure(c(2L,
1L, 1L, 1L, 1L, 3L, 5L, 1L, 1L, 4L, 4L), .Label = c("", "18 vessels",
"98 refrigerated high", "none provided", "seas vessels"), class = "factor"),
company = structure(c(2L, 1L, 1L, 1L, 1L, 1L, 2L, 1L, 1L,
2L, 2L), .Label = c("", "not applicable"), class = "factor"),
area.of.operation = structure(c(2L, 1L, 1L, 1L, 3L, 4L, 2L,
3L, 4L, 2L, 5L), .Label = c("", "above provinces", "annual fishery; EEZ",
"concentrated around", "deepwater coastal"), class = "factor"),
species = structure(c(6L, 3L, 4L, 5L, 9L, 8L, 7L, 9L, 8L,
1L, 2L), .Label = c("Fur seal", "none provided", "otter",
"otter, river", "porpoise", "seal", "Seal", "South American Sea lion,",
"spectacled porpoise,"), class = "factor"), estimates = structure(c(2L,
1L, 1L, 1L, 1L, 1L, 2L, 1L, 1L, 2L, 2L), .Label = c("", "none"
), class = "factor")), class = "data.frame", row.names = c(NA,
-11L))

Answer 1

The old cumsum-split-ting strategem with pasting each column with collapse="," and then sub-ing out the excess commas gets you most of the way:
t( as.data.frame( # transpose because of the column oriented nature of R's apply returns
lapply( split(mydata, cumsum( mydata$target.species != "")),
function(d){ sub(",.*,", ",", lapply( d, paste, collapse=","))})))
[,1] [,2] [,3] [,4] [,5] [,6]
X1 "1,5" "rosefish," "trawl," "18 vessels," "not applicable," "above provinces,annual fishery; EEZ"
X2 "6" "squid," "" "98 refrigerated high" "" "concentrated around"
X3 "7,12" "rosefish," "trawl," "seas vessels," "not applicable," "above provinces,concentrated around"
X4 "13" "hake" "trawl, midwater" "none provided" "not applicable" "above provinces"
X5 "17" "hake, southern" "trawl, bottom" "none provided" "not applicable" "deepwater coastal"
[,7] [,8]
X1 "seal," "none,"
X2 "South American Sea lion," ""
X3 "Seal," "none,"
X4 "Fur seal" "none"
X5 "none provided" "none"